griefith/test
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
Files
00d0a58db37933cd981fb09b057c27cf4c244708
test/source/blender/render/intern/source/rendercore.c

2705 lines
65 KiB
C
Raw Normal View History

Initial revision
2002-10-12 11:37:38 +00:00
/**
* $Id$
*
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
* ***** BEGIN GPL LICENSE BLOCK *****
Initial revision
2002-10-12 11:37:38 +00:00
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
* of the License, or (at your option) any later version.
Initial revision
2002-10-12 11:37:38 +00:00
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*
* The Original Code is Copyright (C) 2001-2002 by NaN Holding BV.
* All rights reserved.
*
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
* Contributors: Hos, Robert Wenzlaff.
* Contributors: 2004/2005/2006 Blender Foundation, full recode
Initial revision
2002-10-12 11:37:38 +00:00
*
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
* ***** END GPL LICENSE BLOCK *****
Initial revision
2002-10-12 11:37:38 +00:00
*/
Bugfix (own discovery!), using sequencer to render a scene updated the depsgraph 2 times to often, also confusing particle drawing when disp % is set. And: little cleanup in render code (move includes in order)
2005-11-23 21:25:20 +00:00
/* system includes */
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
#include <stdio.h>
Bugfix (own discovery!), using sequencer to render a scene updated the depsgraph 2 times to often, also confusing particle drawing when disp % is set. And: little cleanup in render code (move includes in order)
2005-11-23 21:25:20 +00:00
#include <math.h>
Baking now does material texture anti-aliasing if the OSA button is enabled, result of that is mainly visible for image textures.
2008-03-12 20:22:05 +00:00
#include <float.h>
Bugfix (own discovery!), using sequencer to render a scene updated the depsgraph 2 times to often, also confusing particle drawing when disp % is set. And: little cleanup in render code (move includes in order)
2005-11-23 21:25:20 +00:00
#include <string.h>
Initial revision
2002-10-12 11:37:38 +00:00
/* External modules: */
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
#include "MEM_guardedalloc.h"
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
#include "BLI_arithb.h"
#include "BLI_blenlib.h"
#include "BLI_jitter.h"
Thread rendering stability commit. I noticed still several cases where the Imbuf library was called within a thread... and that whilst the Imbuf itself isn't threadsafe. Also the thread lock I added in rendering for loading images actually didn't work, because then it was still possible both threads were accessing the MEM_malloc function at same time. This commit nearly fully replaces ImBuf calls in compositor (giving another nice speedup btw, the way preview images in Nodes were calculated used clumsy imbuf scaling code). I've also centralized the 'mutex' locking for threading, which now only resides in BLI_threads.h. This is used to secure the last ImBuf calls I cannot replace, which is loading images and creating mipmaps. Really hope we get something more stable now!
2006-02-11 15:55:00 +00:00
#include "BLI_rand.h"
#include "BLI_threads.h"
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
Initial revision
2002-10-12 11:37:38 +00:00
#include "BKE_utildefines.h"
Orange branch: Revived hidden treasure, the Groups! Previous experiment (in 2000) didn't satisfy, it had even some primitive NLA option in groups... so, cleaned up the old code (removed most) and integrated it back in a more useful way. Usage: - CTRL+G gives menu to add group, add to existing group, or remove from groups. - In Object buttons, a new (should become first) Panel was added, showing not only Object "ID button" and Parent, but also the Groups the Object Belongs to. These buttons also allow rename, assigning or removing. - To indicate Objects are grouped, they're drawn in a (not theme yet, so temporal?) green wire color. - Use ALT+SHIFT mouse-select to (de)select an entire group But, the real power of groups is in the following features: -> Particle Force field and Guide control In the "Particle Motion" Panel, you can indicate a Group name, this then limits force fields or guides to members of that Group. (Note that layers still work on top of that... not sure about that). -> Light Groups In the Material "Shaders" Panel, you can indicate a Group name to limit lighting for the Material to lamps in this group. The Lights in a Group do need to be 'visible' for the Scene to be rendered (as usual). -> Group Duplicator In the Object "Anim" Panel, you can set any Object (use Empty!) to duplicate an entire Group. It will make copies of all Objects in that Group. Also works for animated Objects, but it will copy the current positions or deforms. Control over 'local timing' (so we can do Massive anims!) will be added later. (Note; this commit won't render Group duplicators yet, a fix in bf-blender will enable that, next commit will sync) -> Library Appending In the SHIFT-F1 or SHIFT+F4 browsers, you can also find the Groups listed. By appending or linking the Group itself, and use the Group Duplicator, you now can animate and position linked Objects. The nice thing is that the local saved file itself will only store the Group name that was linked, so on a next file read, the Group Objects will be re-read as stored (changed) in the Library file. (Note; current implementation also "gives a base" to linked Group Objects, to show them as Objects in the current Scene. Need that now for testing purposes, but probably will be removed later). -> Outliner Outliner now shows Groups as optio too, nice to organize your data a bit too! In General, Groups have a very good potential... for example, it could become default for MetaBall Objects too (jiri, I can help you later on how this works). All current 'layer relationships' in Blender should be dropped in time, I guess...
2005-12-06 10:55:30 +00:00
#include "DNA_image_types.h"
#include "DNA_lamp_types.h"
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
#include "DNA_material_types.h"
Quite a large one this time... but now we have: Edges in Mesh - adds automatic when you use creases. For other situations; call the void make_edges(Mesh *me) in mesh.c. Of course, once in editmode the edges are automatically recreated. - in F9 buttons you can add/remove edges too - both for Mesh and DisplistMesh, so it speeds up drawing quite some in wireframe - render for edges can't work... edges have no material nor tface nor col.. so here still the faces are rendered in wire Creases in Subsurf - based on the code by Chris McFarlen - main changes is that now edges are used, saving quite some data in file - use SHIFT+E in editmode to set edges-sharpness. values go from 0-1 - in F9 buttons you can set draw-crease mode. It draws now blended from wire color to edge-select color (as provided in Theme) Known issue: setting sharpness on 1 cube (subdiv 2) gives weird results with some values... Chris, can you check? Further; code cleanups, changing 0 in NULL when needed, no warnings, etc etc
2004-07-08 20:38:27 +00:00
#include "DNA_meshdata_types.h"
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
#include "DNA_group_types.h"
Initial revision
2002-10-12 11:37:38 +00:00
#include "BKE_global.h"
The Big Image refactor! Please read: http://www.blender3d.org/cms/Imaging.834.0.html Or in short: - adding MultiLayer Image support - recoded entire Image API - better integration of movie/sequence Images Was a whole load of work... went down for a week to do this. So, will need a lot of testing! Will be in irc all evening.
2006-12-20 17:57:56 +00:00
#include "BKE_image.h"
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
#include "BKE_main.h"
Orange: more noodle updates! **** NEW: Group Nodes Node trees usually become messy and confusing quickly, so we need not only a way to collapse Nodes into single 'groups', but also a way to re-use that data to create libraries of effects. This has been done by making a new Library data type, the NodeTree. Everything that has been grouped is stored here, and available for re-use, appending or linking. These NodeTrees are fully generic, i.e. can store shader trees, composit trees, and so on. The 'type' value as stored in the NodeTree will keep track of internal type definitions and execute/drawing callbacks. Needless to say, re-using shader trees in a composit tree is a bit useless, and will be prevented in the browsing code. :) So; any NodeTree can become a "Goup Node" inside in a NodeTree. This Group Node then works just like any Node. To prevent the current code to become too complex, I've disabled the possibility to insert Groups inside of Groups. That might be enabled later, but is a real nasty piece of code to get OK. Since Group Nodes are a dynamic Node type, a lot of work has been done to ensure Node definitions can be dynamic too, but still allow to be stored in files, and allow to be verified for type-definition changes on reloading. This system needs a little bit maturing still, so the Python gurus should better wait a little bit! (Also for me to write the definite API docs for it). What works now: - Press CTRL+G to create a new Group. The grouping code checks for impossible selections (like an unselected node between selected nodes). Everthing that's selected then gets removed from the current tree, and inserted in a new NodeTree library data block. A Group Node then is added which links to this new NodeTree. - Press ALT+G to ungroup. This will not delete the NodeTree library data, but just duplicate the Group into the current tree. - Press TAB, or click on the NodeTree icon to edit Groups. Note that NodeTrees are instances, so editing one Group will also change the other users. This also means that when removing nodes in a Group (or hiding sockets or changing internal links) this is immediately corrected for all users of this Group, also in other Materials. - While editing Groups, only the internal Nodes can be edited. A single click outside of the Group boundary will close this 'edit mode'. What needs to be done: - SHIFT+A menu in toolbox style, also including a list of Groups - Enable the single-user button in the Group Node - Displaying all (visible) internal group UI elements in the Node Panel - Enable Library linking and prevent editing of Groups then. **** NEW: Socket Visibility control Node types will be generated with a lot of possible inputs or outputs, and drawing all sockets all the time isn't very useful then. A new option in the Node header ('plus' icon) allows to either hide all unused sockets (first keypress) or to reveil them (when there are hidden sockets, the icon displays black, otherwise it's blended). Hidden sockets in Nodes also are not exported to a Group, so this way you can control what options (in/outputs) exactly are available. To be done: - a way to hide individual sockets, like with a RMB click on it. **** NEW: Nodes now render! This is still quite primitive, more on a level to replace the (now obsolete and disabled) Material Layers. What needs to be done: - make the "Geometry" node work properly, also for AA textures - make the Texture Node work (does very little at the moment) - give Material Nodes all inputs as needed (like Map-to Panel) - find a way to export more data from a Material Node, like the shadow value, or light intensity only, etc Very important also to separate from the Material Buttons the "global" options, like "Ztransp" or "Wire" or "Halo". These can not be set for each Material-Node individually. Also note that the Preview Render (Buttons window) now renders a bit differently. This was a horrid piece of antique code, using a totally incompatible way of rendering. Target is to fully re-use internal render code for previews. OK... that's it mostly. Now test!
2006-01-02 13:06:05 +00:00
#include "BKE_node.h"
Initial revision
2002-10-12 11:37:38 +00:00
#include "BKE_texture.h"
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
#include "IMB_imbuf_types.h"
#include "IMB_imbuf.h"
Initial revision
2002-10-12 11:37:38 +00:00
/* local include */
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
#include "renderpipeline.h"
#include "render_types.h"
#include "renderdatabase.h"
Approximate Ambient Occlusion ============================= A new approximate ambient occlusion method has been added, next to the existing one based on raytracing. This method is specifically targetted at use in animations, since it is inherently noise free, and so will not flicker across frames. http://www.blender.org/development/current-projects/changes-since-244/approximate-ambient-occlusion/ http://peach.blender.org/index.php/approximate-ambient-occlusion/ Further improvements are still needed, but it can be tested already. There are still a number of known issues: - Bias errors on backfaces. - For performance, instanced object do not occlude currently. - Sky textures don't work well, the derivatives for texture evaluation are not correct. - Multiple passes do not work entirely correct (they are not accurate to begin with, but could be better).
2008-01-17 19:27:16 +00:00
#include "occlusion.h"
Initial revision
2002-10-12 11:37:38 +00:00
#include "pixelblending.h"
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
#include "pixelshading.h"
Initial revision
2002-10-12 11:37:38 +00:00
#include "shadbuf.h"
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
#include "shading.h"
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
#include "sss.h"
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
#include "zbuf.h"
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
#include "RE_raytrace.h"
Initial revision
2002-10-12 11:37:38 +00:00
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
#include "PIL_time.h"
Initial revision
2002-10-12 11:37:38 +00:00
/* own include */
#include "rendercore.h"
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ */
/* defined in pipeline.c, is hardcopy of active dynamic allocated Render */
/* only to be used here in this file, it's for speed */
extern struct Render R;
/* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ */
Initial revision
2002-10-12 11:37:38 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* x and y are current pixels in rect to be rendered */
/* do not normalize! */
Two fixes in one commit. - The ortho render recode had a weak method to calculate the Z value of the render coordinate (shi.co[2]). Worked OK for normal render, but not in unified, giving big errors in Ortho Unfied render. Solved with proper math, using face equation a*x+b*y+c*y=d. Dunno why I didn't do it in first place. :) - Bug fix #2493. Edge render gave much different result for Ortho. This was due to overflows in the integer math filtering zbuffer values. Cleaned it op properly, preventing any integer overflow now.
2005-04-29 08:18:41 +00:00
void calc_view_vector(float *view, float x, float y)
Fixed old annoyance; enabling true Ortho render in Blender. It used to be a simple hack, scaling lens with 100, and moving the camera to the back with an equivalent amount. Because of the hack, making it 100% compatible with older files I could not achieve (yet?). To help reminding users, I've added a print when reading old files with Ortho cameras. Full description of how it works can be found here; http://www.blender3d.com/cms/Render_changes.515.0.html
2005-01-30 11:25:27 +00:00
{
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
Bugfix #5265 3D window, preview panel, didn't give correct specular for Ortho. The view-vector was flipped negative then, so more view-depending code (like tracing) would go wrong then.
2006-11-21 12:41:07 +00:00
view[2]= -ABS(R.clipsta);
Recoded Panorama rendering. The old implementation was added quite hackish (talking about 10 yr ago). You also had to make a small image slice, which was extended Xparts in size. That also required to adjust the camera angle. Very clumsy. Now; when enabling the Panorama option, it will automatically apply the panorama effect on the vertically aligned tiles. You can just enable or disable the "Pano" button, to get a subtle lens effect like this: (without pano) http://www.blender.org/bf/rt.jpg (with pano) http://www.blender.org/bf/rt1.jpg For Panorama render, the minimum slice size has been hardcoded to be 8 pixels. The XParts button goes up to 512 to allow that. In practice, rendering 64 slices will already give very good images for a wide angle lens of 90 degrees, the curvature of straight lines then is equal to a circle of 256 points. Rendering a full 360 degree panorama you do by creating an extreme wide angle camera. The theory says camera-lens 5 should do 360 degrees, but for some reason my tests reveil it's 5.1... there's a rounding error somewhere, maybe related to the clipping plane start? Will look at that later. :) Also note that for each Xpart slice, the entire database needs to be rotated around camera to correct for panorama, on huge scenes that might give some overhead. Threaded render goes fine for Panorama too, but it can only render the vertically aligned parts in parallel. For the next panorama slice it has to wait for all threads of the current slice to be ready. On reading old files, I convert the settings to match as closely as possible the new situation. Since I cannot bump up the version #, the code detects for old panorama by checking for the image size. If image width is smaller than height, it assumes it's an old file (only if Panoroma option was set).
2006-02-27 12:39:36 +00:00
Fixed old annoyance; enabling true Ortho render in Blender. It used to be a simple hack, scaling lens with 100, and moving the camera to the back with an equivalent amount. Because of the hack, making it 100% compatible with older files I could not achieve (yet?). To help reminding users, I've added a print when reading old files with Ortho cameras. Full description of how it works can be found here; http://www.blender3d.com/cms/Render_changes.515.0.html
2005-01-30 11:25:27 +00:00
if(R.r.mode & R_ORTHO) {
Recoded Panorama rendering. The old implementation was added quite hackish (talking about 10 yr ago). You also had to make a small image slice, which was extended Xparts in size. That also required to adjust the camera angle. Very clumsy. Now; when enabling the Panorama option, it will automatically apply the panorama effect on the vertically aligned tiles. You can just enable or disable the "Pano" button, to get a subtle lens effect like this: (without pano) http://www.blender.org/bf/rt.jpg (with pano) http://www.blender.org/bf/rt1.jpg For Panorama render, the minimum slice size has been hardcoded to be 8 pixels. The XParts button goes up to 512 to allow that. In practice, rendering 64 slices will already give very good images for a wide angle lens of 90 degrees, the curvature of straight lines then is equal to a circle of 256 points. Rendering a full 360 degree panorama you do by creating an extreme wide angle camera. The theory says camera-lens 5 should do 360 degrees, but for some reason my tests reveil it's 5.1... there's a rounding error somewhere, maybe related to the clipping plane start? Will look at that later. :) Also note that for each Xpart slice, the entire database needs to be rotated around camera to correct for panorama, on huge scenes that might give some overhead. Threaded render goes fine for Panorama too, but it can only render the vertically aligned parts in parallel. For the next panorama slice it has to wait for all threads of the current slice to be ready. On reading old files, I convert the settings to match as closely as possible the new situation. Since I cannot bump up the version #, the code detects for old panorama by checking for the image size. If image width is smaller than height, it assumes it's an old file (only if Panoroma option was set).
2006-02-27 12:39:36 +00:00
view[0]= view[1]= 0.0f;
Fixed old annoyance; enabling true Ortho render in Blender. It used to be a simple hack, scaling lens with 100, and moving the camera to the back with an equivalent amount. Because of the hack, making it 100% compatible with older files I could not achieve (yet?). To help reminding users, I've added a print when reading old files with Ortho cameras. Full description of how it works can be found here; http://www.blender3d.com/cms/Render_changes.515.0.html
2005-01-30 11:25:27 +00:00
}
else {
Recoded Panorama rendering. The old implementation was added quite hackish (talking about 10 yr ago). You also had to make a small image slice, which was extended Xparts in size. That also required to adjust the camera angle. Very clumsy. Now; when enabling the Panorama option, it will automatically apply the panorama effect on the vertically aligned tiles. You can just enable or disable the "Pano" button, to get a subtle lens effect like this: (without pano) http://www.blender.org/bf/rt.jpg (with pano) http://www.blender.org/bf/rt1.jpg For Panorama render, the minimum slice size has been hardcoded to be 8 pixels. The XParts button goes up to 512 to allow that. In practice, rendering 64 slices will already give very good images for a wide angle lens of 90 degrees, the curvature of straight lines then is equal to a circle of 256 points. Rendering a full 360 degree panorama you do by creating an extreme wide angle camera. The theory says camera-lens 5 should do 360 degrees, but for some reason my tests reveil it's 5.1... there's a rounding error somewhere, maybe related to the clipping plane start? Will look at that later. :) Also note that for each Xpart slice, the entire database needs to be rotated around camera to correct for panorama, on huge scenes that might give some overhead. Threaded render goes fine for Panorama too, but it can only render the vertically aligned parts in parallel. For the next panorama slice it has to wait for all threads of the current slice to be ready. On reading old files, I convert the settings to match as closely as possible the new situation. Since I cannot bump up the version #, the code detects for old panorama by checking for the image size. If image width is smaller than height, it assumes it's an old file (only if Panoroma option was set).
2006-02-27 12:39:36 +00:00
if(R.r.mode & R_PANORAMA)
x-= R.panodxp;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* move x and y to real viewplane coords */
x= (x/(float)R.winx);
view[0]= R.viewplane.xmin + x*(R.viewplane.xmax - R.viewplane.xmin);
Fixed old annoyance; enabling true Ortho render in Blender. It used to be a simple hack, scaling lens with 100, and moving the camera to the back with an equivalent amount. Because of the hack, making it 100% compatible with older files I could not achieve (yet?). To help reminding users, I've added a print when reading old files with Ortho cameras. Full description of how it works can be found here; http://www.blender3d.com/cms/Render_changes.515.0.html
2005-01-30 11:25:27 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
y= (y/(float)R.winy);
view[1]= R.viewplane.ymin + y*(R.viewplane.ymax - R.viewplane.ymin);
// if(R.flag & R_SEC_FIELD) {
// if(R.r.mode & R_ODDFIELD) view[1]= (y+R.ystart)*R.ycor;
// else view[1]= (y+R.ystart+1.0)*R.ycor;
// }
// else view[1]= (y+R.ystart+R.bluroffsy+0.5)*R.ycor;
Recoded Panorama rendering. The old implementation was added quite hackish (talking about 10 yr ago). You also had to make a small image slice, which was extended Xparts in size. That also required to adjust the camera angle. Very clumsy. Now; when enabling the Panorama option, it will automatically apply the panorama effect on the vertically aligned tiles. You can just enable or disable the "Pano" button, to get a subtle lens effect like this: (without pano) http://www.blender.org/bf/rt.jpg (with pano) http://www.blender.org/bf/rt1.jpg For Panorama render, the minimum slice size has been hardcoded to be 8 pixels. The XParts button goes up to 512 to allow that. In practice, rendering 64 slices will already give very good images for a wide angle lens of 90 degrees, the curvature of straight lines then is equal to a circle of 256 points. Rendering a full 360 degree panorama you do by creating an extreme wide angle camera. The theory says camera-lens 5 should do 360 degrees, but for some reason my tests reveil it's 5.1... there's a rounding error somewhere, maybe related to the clipping plane start? Will look at that later. :) Also note that for each Xpart slice, the entire database needs to be rotated around camera to correct for panorama, on huge scenes that might give some overhead. Threaded render goes fine for Panorama too, but it can only render the vertically aligned parts in parallel. For the next panorama slice it has to wait for all threads of the current slice to be ready. On reading old files, I convert the settings to match as closely as possible the new situation. Since I cannot bump up the version #, the code detects for old panorama by checking for the image size. If image width is smaller than height, it assumes it's an old file (only if Panoroma option was set).
2006-02-27 12:39:36 +00:00
if(R.r.mode & R_PANORAMA) {
float u= view[0] + R.panodxv; float v= view[2];
view[0]= R.panoco*u + R.panosi*v;
view[2]= -R.panosi*u + R.panoco*v;
}
Initial revision
2002-10-12 11:37:38 +00:00
}
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
void calc_renderco_ortho(float *co, float x, float y, int z)
Initial revision
2002-10-12 11:37:38 +00:00
{
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* x and y 3d coordinate can be derived from pixel coord and winmat */
float fx= 2.0f/(R.winx*R.winmat[0][0]);
float fy= 2.0f/(R.winy*R.winmat[1][1]);
float zco;
Initial revision
2002-10-12 11:37:38 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
co[0]= (x - 0.5f*R.winx)*fx - R.winmat[3][0]/R.winmat[0][0];
co[1]= (y - 0.5f*R.winy)*fy - R.winmat[3][1]/R.winmat[1][1];
Initial revision
2002-10-12 11:37:38 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
zco= ((float)z)/2147483647.0f;
co[2]= R.winmat[3][2]/( R.winmat[2][3]*zco - R.winmat[2][2] );
Initial revision
2002-10-12 11:37:38 +00:00
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
void calc_renderco_zbuf(float *co, float *view, int z)
Initial revision
2002-10-12 11:37:38 +00:00
{
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
float fac, zco;
Initial revision
2002-10-12 11:37:38 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* inverse of zbuf calc: zbuf = MAXZ*hoco_z/hoco_w */
zco= ((float)z)/2147483647.0f;
co[2]= R.winmat[3][2]/( R.winmat[2][3]*zco - R.winmat[2][2] );
Bugfix #2943 Render bug. Alpha low or zero, Ztransp, didnt render with lamp halos over... Note it still doesnt work for ray-transp!
2005-08-14 19:26:58 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
fac= co[2]/view[2];
co[0]= fac*view[0];
co[1]= fac*view[1];
Initial revision
2002-10-12 11:37:38 +00:00
}
New shadow feature: Irregular Shadow Buffers Full log: http://www.blender3d.org/cms/Irregular_Shadow_Buffe.785.0.html In short: this is a shadow buffer approach that always results in crispy shadows, independent of lamp buffer size or zoom level. This shadow buffer system also supports transparent shadow. This is part of work on refreshing Shadow Buffers in Blender. You now can choose of two types (Classical, Irregular). More types will follow. Also quality issues for Classical shadow buffers are going to be reviewed, especially to solve the lousy Biasing. For the CVS log record; it is based on articles: Gregory Johnson et al, University of Texas, Austin. (Regular grid method). Timo Aila and Samuli Laine, Helsinki University of Technology. (BSP method).
2006-10-14 10:21:19 +00:00
/* also used in zbuf.c and shadbuf.c */
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
int count_mask(unsigned short mask)
Initial revision
2002-10-12 11:37:38 +00:00
{
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
if(R.samples)
return (R.samples->cmask[mask & 255]+R.samples->cmask[mask>>8]);
return 0;
}
Initial revision
2002-10-12 11:37:38 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
static int calchalo_z(HaloRen *har, int zz)
{
Initial revision
2002-10-12 11:37:38 +00:00
if(har->type & HA_ONLYSKY) {
Bugfix #13675 Edge render made star render disappear. Wrong Z value comparing... (2 year old bug!)
2008-09-09 11:40:29 +00:00
if(zz < 0x7FFFFFF0) zz= - 0x7FFFFF; /* edge render messes zvalues */
Initial revision
2002-10-12 11:37:38 +00:00
}
else {
zz= (zz>>8);
}
return zz;
}
Patch #8034: "soft" option for halos, which avoids ugly intersections with geometry, and makes halos look more volumetric. Patch contributed by Markus Ilmola, thanks!
2008-04-14 19:48:14 +00:00
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
static void halo_pixelstruct(HaloRen *har, RenderLayer **rlpp, int totsample, int od, float dist, float xn, float yn, PixStr *ps)
Initial revision
2002-10-12 11:37:38 +00:00
{
Patch #8034: "soft" option for halos, which avoids ugly intersections with geometry, and makes halos look more volumetric. Patch contributed by Markus Ilmola, thanks!
2008-04-14 19:48:14 +00:00
float col[4], accol[4], fac;
Fix for bug #8927: halo blending with solid has poor antialiasing, which as far as I can see is an old issue, but with FSA it is quite simple to do better.
2008-04-12 21:44:08 +00:00
int amount, amountm, zz, flarec, sample, fullsample, mask=0;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
Fix for bug #8927: halo blending with solid has poor antialiasing, which as far as I can see is an old issue, but with FSA it is quite simple to do better.
2008-04-12 21:44:08 +00:00
fullsample= (totsample > 1);
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
amount= 0;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
accol[0]=accol[1]=accol[2]=accol[3]= 0.0f;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
flarec= har->flarec;
while(ps) {
amountm= count_mask(ps->mask);
amount+= amountm;
zz= calchalo_z(har, ps->z);
Fix for bug #9075: the crash with stars rendering was not completely fixed yet.
2008-04-17 12:11:15 +00:00
if((zz> har->zs) || (har->mat && (har->mat->mode & MA_HALO_SOFT))) {
Patch #8034: "soft" option for halos, which avoids ugly intersections with geometry, and makes halos look more volumetric. Patch contributed by Markus Ilmola, thanks!
2008-04-14 19:48:14 +00:00
if(shadeHaloFloat(har, col, zz, dist, xn, yn, flarec)) {
flarec= 0;
if(fullsample) {
for(sample=0; sample<totsample; sample++)
if(ps->mask & (1 << sample))
addalphaAddfacFloat(rlpp[sample]->rectf + od*4, col, har->add);
}
else {
fac= ((float)amountm)/(float)R.osa;
accol[0]+= fac*col[0];
accol[1]+= fac*col[1];
accol[2]+= fac*col[2];
accol[3]+= fac*col[3];
}
Fix for bug #8927: halo blending with solid has poor antialiasing, which as far as I can see is an old issue, but with FSA it is quite simple to do better.
2008-04-12 21:44:08 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
}
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
Fix for bug #8927: halo blending with solid has poor antialiasing, which as far as I can see is an old issue, but with FSA it is quite simple to do better.
2008-04-12 21:44:08 +00:00
mask |= ps->mask;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
ps= ps->next;
Initial revision
2002-10-12 11:37:38 +00:00
}
Fix for bug #8927: halo blending with solid has poor antialiasing, which as far as I can see is an old issue, but with FSA it is quite simple to do better.
2008-04-12 21:44:08 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* now do the sky sub-pixels */
amount= R.osa-amount;
if(amount) {
Patch #8034: "soft" option for halos, which avoids ugly intersections with geometry, and makes halos look more volumetric. Patch contributed by Markus Ilmola, thanks!
2008-04-14 19:48:14 +00:00
if(shadeHaloFloat(har, col, 0x7FFFFF, dist, xn, yn, flarec)) {
if(!fullsample) {
fac= ((float)amount)/(float)R.osa;
accol[0]+= fac*col[0];
accol[1]+= fac*col[1];
accol[2]+= fac*col[2];
accol[3]+= fac*col[3];
}
Fix for bug #8927: halo blending with solid has poor antialiasing, which as far as I can see is an old issue, but with FSA it is quite simple to do better.
2008-04-12 21:44:08 +00:00
}
}
if(fullsample) {
for(sample=0; sample<totsample; sample++)
if(!(mask & (1 << sample)))
addalphaAddfacFloat(rlpp[sample]->rectf + od*4, col, har->add);
}
else {
col[0]= accol[0];
col[1]= accol[1];
col[2]= accol[2];
col[3]= accol[3];
for(sample=0; sample<totsample; sample++)
addalphaAddfacFloat(rlpp[sample]->rectf + od*4, col, har->add);
Another round of cleanups for render code; - removed ugly pointerhack from OSA render (negative indices denoted pointers). this should solve memory errors when using >1.5 gig mem - cleaned up usage of zbuffer values. These are signed in Blender, and treated as unsigned all over, giving confusing code - fixed incorrect gamma-adding for halos (caused in after xmas commit) And bugfix #2101; wire render didn't give correct rendering for mist. This caused by fact wires are 2D pixel lines, and not correctly filled in faces. Retrieving the 3d coordinate while render cannot use a face- equation then. Solved by retrieving 3D coordinate based on zbuffer value. Still todo here: calculating correct texture coordinates for wire-edges that are no faces.
2005-01-07 14:11:00 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
}
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
static void halo_tile(RenderPart *pa, RenderLayer *rl)
Initial revision
2002-10-12 11:37:38 +00:00
{
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
RenderLayer *rlpp[RE_MAX_OSA];
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
HaloRen *har;
Bugfix #4454 Lens flare: error in counting visibility per halo-flare in the to-be cropped area of tiles. Halos now don't render in the 2 pixels exta outline per tile, which is only used by faces anyway.
2006-06-25 10:52:13 +00:00
rcti disprect= pa->disprect, testrect= pa->disprect;
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
float dist, xsq, ysq, xn, yn;
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
float col[4];
Win64: please check my changes if you ran across them ;) But should be fine since no additional crashes were reported!
2008-08-17 17:08:00 +00:00
intptr_t *rd= NULL;
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
int a, *rz, zz, y, sample, totsample, od;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
short minx, maxx, miny, maxy, x;
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
unsigned int lay= rl->lay;
Initial revision
2002-10-12 11:37:38 +00:00
Bugfix #4454 Lens flare: error in counting visibility per halo-flare in the to-be cropped area of tiles. Halos now don't render in the 2 pixels exta outline per tile, which is only used by faces anyway.
2006-06-25 10:52:13 +00:00
/* we don't render halos in the cropped area, gives errors in flare counter */
if(pa->crop) {
testrect.xmin+= pa->crop;
testrect.xmax-= pa->crop;
testrect.ymin+= pa->crop;
testrect.ymax-= pa->crop;
}
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
totsample= get_sample_layers(pa, rl, rlpp);
Initial revision
2002-10-12 11:37:38 +00:00
for(a=0; a<R.tothalo; a++) {
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
har= R.sortedhalos[a];
Initial revision
2002-10-12 11:37:38 +00:00
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
/* layer test, clip halo with y */
if((har->lay & lay)==0);
Bugfix #4454 Lens flare: error in counting visibility per halo-flare in the to-be cropped area of tiles. Halos now don't render in the 2 pixels exta outline per tile, which is only used by faces anyway.
2006-06-25 10:52:13 +00:00
else if(testrect.ymin > har->maxy);
else if(testrect.ymax < har->miny);
Initial revision
2002-10-12 11:37:38 +00:00
else {
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
Initial revision
2002-10-12 11:37:38 +00:00
minx= floor(har->xs-har->rad);
maxx= ceil(har->xs+har->rad);
Bugfix #4454 Lens flare: error in counting visibility per halo-flare in the to-be cropped area of tiles. Halos now don't render in the 2 pixels exta outline per tile, which is only used by faces anyway.
2006-06-25 10:52:13 +00:00
if(testrect.xmin > maxx);
else if(testrect.xmax < minx);
Initial revision
2002-10-12 11:37:38 +00:00
else {
Bugfix #4454 Lens flare: error in counting visibility per halo-flare in the to-be cropped area of tiles. Halos now don't render in the 2 pixels exta outline per tile, which is only used by faces anyway.
2006-06-25 10:52:13 +00:00
minx= MAX2(minx, testrect.xmin);
maxx= MIN2(maxx, testrect.xmax);
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
Bugfix #4454 Lens flare: error in counting visibility per halo-flare in the to-be cropped area of tiles. Halos now don't render in the 2 pixels exta outline per tile, which is only used by faces anyway.
2006-06-25 10:52:13 +00:00
miny= MAX2(har->miny, testrect.ymin);
maxy= MIN2(har->maxy, testrect.ymax);
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
for(y=miny; y<maxy; y++) {
int rectofs= (y-disprect.ymin)*pa->rectx + (minx - disprect.xmin);
rz= pa->rectz + rectofs;
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
od= rectofs;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
if(pa->rectdaps)
rd= pa->rectdaps + rectofs;
yn= (y-har->ys)*R.ycor;
ysq= yn*yn;
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
for(x=minx; x<maxx; x++, rz++, od++) {
Initial revision
2002-10-12 11:37:38 +00:00
xn= x- har->xs;
xsq= xn*xn;
dist= xsq+ysq;
if(dist<har->radsq) {
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
if(rd && *rd) {
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
halo_pixelstruct(har, rlpp, totsample, od, dist, xn, yn, (PixStr *)*rd);
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
}
else {
zz= calchalo_z(har, *rz);
Fix for bug #9075: the crash with stars rendering was not completely fixed yet.
2008-04-17 12:11:15 +00:00
if((zz> har->zs) || (har->mat && (har->mat->mode & MA_HALO_SOFT))) {
Patch #8034: "soft" option for halos, which avoids ugly intersections with geometry, and makes halos look more volumetric. Patch contributed by Markus Ilmola, thanks!
2008-04-14 19:48:14 +00:00
if(shadeHaloFloat(har, col, zz, dist, xn, yn, har->flarec)) {
for(sample=0; sample<totsample; sample++)
addalphaAddfacFloat(rlpp[sample]->rectf + od*4, col, har->add);
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
}
}
}
if(rd) rd++;
}
}
Initial revision
2002-10-12 11:37:38 +00:00
}
}
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(R.test_break(R.tbh) ) break;
Initial revision
2002-10-12 11:37:38 +00:00
}
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
Numerous fixes in Render code: - Bug: material emit was ignored (showed in preview render backdrop) - Bug: world exposure was ignored - Bug: lamp halo was ignoring 'render layer light override'. Further reshuffled the way shadows are being pre-calculated, this to enable more advanced (and faster) usage of Material lightgroups. Now shadows are being cached in lamps, using a per-sample counter to check if a recalc is needed. Will also work (later) for Raytracing node shaders. - New: Material LightGroup option "Always", which always shades the lights in the group, independent of visibility layer. (so it allows to move such lights to hidden layer, not influencing anything).
2006-12-08 09:40:44 +00:00
static void lamphalo_tile(RenderPart *pa, RenderLayer *rl)
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
{
Fix for bug #8910: FSA + spotlight halo not working correct together.
2008-04-29 19:51:23 +00:00
RenderLayer *rlpp[RE_MAX_OSA];
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
ShadeInput shi;
Fix for bug #8910: FSA + spotlight halo not working correct together.
2008-04-29 19:51:23 +00:00
float *pass;
float fac, col[4];
Win64: please check my changes if you ran across them ;) But should be fine since no additional crashes were reported!
2008-08-17 17:08:00 +00:00
intptr_t *rd= pa->rectdaps;
Fix for bug #8910: FSA + spotlight halo not working correct together.
2008-04-29 19:51:23 +00:00
int *rz= pa->rectz;
int x, y, sample, totsample, fullsample, od;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
Fix for bug #8910: FSA + spotlight halo not working correct together.
2008-04-29 19:51:23 +00:00
totsample= get_sample_layers(pa, rl, rlpp);
fullsample= (totsample > 1);
Bugfix [#6861] Black dots when using small lamps on Mirror materials in 2.44 and 2.43. some values were not initialized properly, for example, the window coordinates for reflections, this caused NAN color values for some pixels, (may also fix plumiferos bad pixel problem from last bconf)
2008-01-13 11:15:23 +00:00
shade_input_initialize(&shi, pa, rl, 0); /* this zero's ShadeInput for us */
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
Fix for bug #8910: FSA + spotlight halo not working correct together.
2008-04-29 19:51:23 +00:00
for(od=0, y=pa->disprect.ymin; y<pa->disprect.ymax; y++) {
for(x=pa->disprect.xmin; x<pa->disprect.xmax; x++, rz++, od++) {
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
calc_view_vector(shi.view, x, y);
if(rd && *rd) {
PixStr *ps= (PixStr *)*rd;
Fix for bug #8910: FSA + spotlight halo not working correct together.
2008-04-29 19:51:23 +00:00
int count, totsamp= 0, mask= 0;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
while(ps) {
if(R.r.mode & R_ORTHO)
calc_renderco_ortho(shi.co, (float)x, (float)y, ps->z);
else
calc_renderco_zbuf(shi.co, shi.view, ps->z);
Fix for bug #8910: FSA + spotlight halo not working correct together.
2008-04-29 19:51:23 +00:00
totsamp+= count= count_mask(ps->mask);
mask |= ps->mask;
col[0]= col[1]= col[2]= col[3]= 0.0f;
renderspothalo(&shi, col, 1.0f);
if(fullsample) {
for(sample=0; sample<totsample; sample++) {
if(ps->mask & (1 << sample)) {
pass= rlpp[sample]->rectf + od*4;
pass[0]+= col[0];
pass[1]+= col[1];
pass[2]+= col[2];
pass[3]+= col[3];
if(pass[3]>1.0f) pass[3]= 1.0f;
}
}
}
else {
fac= ((float)count)/(float)R.osa;
pass= rl->rectf + od*4;
pass[0]+= fac*col[0];
pass[1]+= fac*col[1];
pass[2]+= fac*col[2];
pass[3]+= fac*col[3];
if(pass[3]>1.0f) pass[3]= 1.0f;
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
ps= ps->next;
}
Fix for bug #8910: FSA + spotlight halo not working correct together.
2008-04-29 19:51:23 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
if(totsamp<R.osa) {
shi.co[2]= 0.0f;
Fix for bug #8910: FSA + spotlight halo not working correct together.
2008-04-29 19:51:23 +00:00
col[0]= col[1]= col[2]= col[3]= 0.0f;
renderspothalo(&shi, col, 1.0f);
if(fullsample) {
for(sample=0; sample<totsample; sample++) {
if(!(mask & (1 << sample))) {
pass= rlpp[sample]->rectf + od*4;
pass[0]+= col[0];
pass[1]+= col[1];
pass[2]+= col[2];
pass[3]+= col[3];
if(pass[3]>1.0f) pass[3]= 1.0f;
}
}
}
else {
fac= ((float)R.osa-totsamp)/(float)R.osa;
pass= rl->rectf + od*4;
pass[0]+= fac*col[0];
pass[1]+= fac*col[1];
pass[2]+= fac*col[2];
pass[3]+= fac*col[3];
if(pass[3]>1.0f) pass[3]= 1.0f;
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
}
}
else {
if(R.r.mode & R_ORTHO)
calc_renderco_ortho(shi.co, (float)x, (float)y, *rz);
else
calc_renderco_zbuf(shi.co, shi.view, *rz);
Fix for bug #8910: FSA + spotlight halo not working correct together.
2008-04-29 19:51:23 +00:00
col[0]= col[1]= col[2]= col[3]= 0.0f;
renderspothalo(&shi, col, 1.0f);
for(sample=0; sample<totsample; sample++) {
pass= rlpp[sample]->rectf + od*4;
pass[0]+= col[0];
pass[1]+= col[1];
pass[2]+= col[2];
pass[3]+= col[3];
if(pass[3]>1.0f) pass[3]= 1.0f;
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
}
if(rd) rd++;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
if(y&1)
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(R.test_break(R.tbh)) break;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
first commit with hours and hours of comment translation for render code...
2003-04-27 14:05:41 +00:00
/* ********************* MAINLOOPS ******************** */
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
/* osa version */
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
static void add_filt_passes(RenderLayer *rl, int curmask, int rectx, int offset, ShadeInput *shi, ShadeResult *shr)
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
{
RenderPass *rpass;
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
/* combined rgb */
add_filt_fmask(curmask, shr->combined, rl->rectf + 4*offset, rectx);
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
for(rpass= rl->passes.first; rpass; rpass= rpass->next) {
float *fp, *col= NULL;
int pixsize= 3;
switch(rpass->passtype) {
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
case SCE_PASS_Z:
fp= rpass->rect + offset;
- Bugfix for ztransp not being filled into the z-pass after FSA changes. - Bugfix for speedvectors being not properly set for some strands.
2008-01-30 13:35:28 +00:00
*fp= shr->z;
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
break;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
case SCE_PASS_RGBA:
col= shr->col;
pixsize= 4;
break;
case SCE_PASS_DIFFUSE:
Work on RenderLayer and Pass control: Full log: http://www.blender3d.org/cms/Render_Passes.829.0.html In short: - Passes now have option to be excluded from "Combined". - RenderLayers allow to override Light (Lamp groups) or Material. - RenderLayers and Passes are in Outliner now, (ab)using Matt's nice 'restriction collumns'. :)
2006-12-07 14:17:38 +00:00
col= shr->diff;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
break;
case SCE_PASS_SPEC:
col= shr->spec;
break;
case SCE_PASS_SHADOW:
col= shr->shad;
break;
case SCE_PASS_AO:
col= shr->ao;
break;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
case SCE_PASS_REFLECT:
col= shr->refl;
break;
case SCE_PASS_REFRACT:
col= shr->refr;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
break;
Composite & Pass render goodies: - New Passes: UV and Rad(iosity) - New Nodes: UV Map and Index Mask - Z-combine now is antialiased As usual, please check the log. Has nice pics! http://www.blender3d.org/cms/Composite__UV_Map__ID.830.0.html For devs: the antialias code from Vector Blur is now exported in compo too. Works pretty good. Even fixed a bug in antialias, so vectorblur will be better. Also: found out that OpenGL display list speedup accidentally was still triggered with the rt button... so it did not work by default.
2006-12-10 20:30:15 +00:00
case SCE_PASS_RADIO:
col= shr->rad;
break;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
case SCE_PASS_NORMAL:
col= shr->nor;
break;
Composite & Pass render goodies: - New Passes: UV and Rad(iosity) - New Nodes: UV Map and Index Mask - Z-combine now is antialiased As usual, please check the log. Has nice pics! http://www.blender3d.org/cms/Composite__UV_Map__ID.830.0.html For devs: the antialias code from Vector Blur is now exported in compo too. Works pretty good. Even fixed a bug in antialias, so vectorblur will be better. Also: found out that OpenGL display list speedup accidentally was still triggered with the rt button... so it did not work by default.
2006-12-10 20:30:15 +00:00
case SCE_PASS_UV:
/* box filter only, gauss will screwup UV too much */
Bugfix #5712 UV rendering error: the third value of a UV texture coordinate should be set to zero always, this enables 3d textures to work uniform with UV. Also fixed: UV pass output for non_OSA is now equal to OSA, meaning only used UV face values are in such passes.
2007-01-15 11:41:10 +00:00
if(shi->totuv) {
Composite & Pass render goodies: - New Passes: UV and Rad(iosity) - New Nodes: UV Map and Index Mask - Z-combine now is antialiased As usual, please check the log. Has nice pics! http://www.blender3d.org/cms/Composite__UV_Map__ID.830.0.html For devs: the antialias code from Vector Blur is now exported in compo too. Works pretty good. Even fixed a bug in antialias, so vectorblur will be better. Also: found out that OpenGL display list speedup accidentally was still triggered with the rt button... so it did not work by default.
2006-12-10 20:30:15 +00:00
float mult= (float)count_mask(curmask)/(float)R.osa;
fp= rpass->rect + 3*offset;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
fp[0]+= mult*(0.5f + 0.5f*shi->uv[shi->actuv].uv[0]);
fp[1]+= mult*(0.5f + 0.5f*shi->uv[shi->actuv].uv[1]);
Composite & Pass render goodies: - New Passes: UV and Rad(iosity) - New Nodes: UV Map and Index Mask - Z-combine now is antialiased As usual, please check the log. Has nice pics! http://www.blender3d.org/cms/Composite__UV_Map__ID.830.0.html For devs: the antialias code from Vector Blur is now exported in compo too. Works pretty good. Even fixed a bug in antialias, so vectorblur will be better. Also: found out that OpenGL display list speedup accidentally was still triggered with the rt button... so it did not work by default.
2006-12-10 20:30:15 +00:00
fp[2]+= mult;
}
break;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
case SCE_PASS_INDEXOB:
/* no filter */
if(shi->vlr) {
fp= rpass->rect + offset;
Composite & Pass render goodies: - New Passes: UV and Rad(iosity) - New Nodes: UV Map and Index Mask - Z-combine now is antialiased As usual, please check the log. Has nice pics! http://www.blender3d.org/cms/Composite__UV_Map__ID.830.0.html For devs: the antialias code from Vector Blur is now exported in compo too. Works pretty good. Even fixed a bug in antialias, so vectorblur will be better. Also: found out that OpenGL display list speedup accidentally was still triggered with the rt button... so it did not work by default.
2006-12-10 20:30:15 +00:00
if(*fp==0.0f)
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
*fp= (float)shi->obr->ob->index;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
}
break;
Added new render pass: "Mist". This is actually just the alpha value as currently being calculated by the mist code. It is in many cases not very useful to have this as alpha in shading result, also for postprocess and composite. Note: this pass also works with "Mist" not set in World, of course.
2008-01-19 11:17:12 +00:00
case SCE_PASS_MIST:
/* */
col= &shr->mist;
pixsize= 1;
break;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
case SCE_PASS_VECTOR:
iImage based Vector Blur After a couple of experiments with variable blur filters, I tried a more interesting, and who knows... original approach. :) First watch results here: http://www.blender.org/bf/rt0001_0030.avi http://www.blender.org/bf/hand0001_0060.avi These are the steps in producing such results: - In preprocess, the speed vectors to previous and next frame are calculated. Speed vectors are screen-aligned and in pixel size. - while rendering, these vectors get calculated per sample, and accumulated in the vector buffer checking for "minimum speed". (on start the vector buffer is initialized on max speed). - After render: - The entire image, all pixels, then is converted to quad polygons. - Also the z value of the pixels is assigned to the polygons - The vertices for the quads use averaged speed vectors (of the 4 corner faces), using a 'minimum but non-zero' speed rule. This minimal speed trick works very well to prevent 'tearing' apart when multiple faces move in different directions in a pixel, or to be able to separate moving pixels clearly from non-moving ones - So, now we have a sort of 'mask' of quad polygons. The previous steps guaranteed that this mask doesn't have antialias color info, and has speed vectors that ensure individual parts to move nicely without tearing effects. The Z allows multiple layers of moving masks. - Then, in temporal buffer, faces get tagged if they move or not - These tags then go to an anti-alias routine, which assigns alpha values to edge faces, based on the method we used in past to antialias bitmaps (still in our code, check the antialias.c in imbuf!) - finally, the tag buffer is used to tag which z values of the original image have to be included (to allow blur go behind stuff). - OK, now we're ready for accumulating! In a loop, all faces then get drawn (with zbuffer) with increasing influence of their speed vectors. The resulting image then is accumulated on top of the original with a decreasing weighting value. It sounds all quite complex... but the speed is still encouraging. Above images have 64 mblur steps, which takes about 1-3 seconds per frame. Usage notes: - Make sure the render-layer has passes 'Vector' and 'Z' on. - add in Compositor the VectorBlur node, and connect the image, Z and speed to the inputs. - The node allows to set amount of steps (10 steps = 10 forward, 10 back). and to set a maximum speed in pixels... to prevent extreme moving things to blur too wide.
2006-02-06 22:11:50 +00:00
{
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* add minimum speed in pixel, no filter */
iImage based Vector Blur After a couple of experiments with variable blur filters, I tried a more interesting, and who knows... original approach. :) First watch results here: http://www.blender.org/bf/rt0001_0030.avi http://www.blender.org/bf/hand0001_0060.avi These are the steps in producing such results: - In preprocess, the speed vectors to previous and next frame are calculated. Speed vectors are screen-aligned and in pixel size. - while rendering, these vectors get calculated per sample, and accumulated in the vector buffer checking for "minimum speed". (on start the vector buffer is initialized on max speed). - After render: - The entire image, all pixels, then is converted to quad polygons. - Also the z value of the pixels is assigned to the polygons - The vertices for the quads use averaged speed vectors (of the 4 corner faces), using a 'minimum but non-zero' speed rule. This minimal speed trick works very well to prevent 'tearing' apart when multiple faces move in different directions in a pixel, or to be able to separate moving pixels clearly from non-moving ones - So, now we have a sort of 'mask' of quad polygons. The previous steps guaranteed that this mask doesn't have antialias color info, and has speed vectors that ensure individual parts to move nicely without tearing effects. The Z allows multiple layers of moving masks. - Then, in temporal buffer, faces get tagged if they move or not - These tags then go to an anti-alias routine, which assigns alpha values to edge faces, based on the method we used in past to antialias bitmaps (still in our code, check the antialias.c in imbuf!) - finally, the tag buffer is used to tag which z values of the original image have to be included (to allow blur go behind stuff). - OK, now we're ready for accumulating! In a loop, all faces then get drawn (with zbuffer) with increasing influence of their speed vectors. The resulting image then is accumulated on top of the original with a decreasing weighting value. It sounds all quite complex... but the speed is still encouraging. Above images have 64 mblur steps, which takes about 1-3 seconds per frame. Usage notes: - Make sure the render-layer has passes 'Vector' and 'Z' on. - add in Compositor the VectorBlur node, and connect the image, Z and speed to the inputs. - The node allows to set amount of steps (10 steps = 10 forward, 10 back). and to set a maximum speed in pixels... to prevent extreme moving things to blur too wide.
2006-02-06 22:11:50 +00:00
fp= rpass->rect + 4*offset;
if( (ABS(shr->winspeed[0]) + ABS(shr->winspeed[1]))< (ABS(fp[0]) + ABS(fp[1])) ) {
fp[0]= shr->winspeed[0];
fp[1]= shr->winspeed[1];
}
if( (ABS(shr->winspeed[2]) + ABS(shr->winspeed[3]))< (ABS(fp[2]) + ABS(fp[3])) ) {
fp[2]= shr->winspeed[2];
fp[3]= shr->winspeed[3];
}
}
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
break;
}
if(col) {
fp= rpass->rect + pixsize*offset;
add_filt_fmask_pixsize(curmask, col, fp, rectx, pixsize);
}
}
}
/* non-osa version */
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
static void add_passes(RenderLayer *rl, int offset, ShadeInput *shi, ShadeResult *shr)
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
{
RenderPass *rpass;
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
float *fp;
fp= rl->rectf + 4*offset;
QUATCOPY(fp, shr->combined);
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
for(rpass= rl->passes.first; rpass; rpass= rpass->next) {
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
float *col= NULL, uvcol[3];
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
int a, pixsize= 3;
switch(rpass->passtype) {
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
case SCE_PASS_Z:
fp= rpass->rect + offset;
- Bugfix for ztransp not being filled into the z-pass after FSA changes. - Bugfix for speedvectors being not properly set for some strands.
2008-01-30 13:35:28 +00:00
*fp= shr->z;
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
break;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
case SCE_PASS_RGBA:
col= shr->col;
pixsize= 4;
break;
case SCE_PASS_DIFFUSE:
Work on RenderLayer and Pass control: Full log: http://www.blender3d.org/cms/Render_Passes.829.0.html In short: - Passes now have option to be excluded from "Combined". - RenderLayers allow to override Light (Lamp groups) or Material. - RenderLayers and Passes are in Outliner now, (ab)using Matt's nice 'restriction collumns'. :)
2006-12-07 14:17:38 +00:00
col= shr->diff;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
break;
case SCE_PASS_SPEC:
col= shr->spec;
break;
case SCE_PASS_SHADOW:
col= shr->shad;
break;
case SCE_PASS_AO:
col= shr->ao;
break;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
case SCE_PASS_REFLECT:
col= shr->refl;
break;
case SCE_PASS_REFRACT:
col= shr->refr;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
break;
Composite & Pass render goodies: - New Passes: UV and Rad(iosity) - New Nodes: UV Map and Index Mask - Z-combine now is antialiased As usual, please check the log. Has nice pics! http://www.blender3d.org/cms/Composite__UV_Map__ID.830.0.html For devs: the antialias code from Vector Blur is now exported in compo too. Works pretty good. Even fixed a bug in antialias, so vectorblur will be better. Also: found out that OpenGL display list speedup accidentally was still triggered with the rt button... so it did not work by default.
2006-12-10 20:30:15 +00:00
case SCE_PASS_RADIO:
col= shr->rad;
break;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
case SCE_PASS_NORMAL:
col= shr->nor;
break;
Composite & Pass render goodies: - New Passes: UV and Rad(iosity) - New Nodes: UV Map and Index Mask - Z-combine now is antialiased As usual, please check the log. Has nice pics! http://www.blender3d.org/cms/Composite__UV_Map__ID.830.0.html For devs: the antialias code from Vector Blur is now exported in compo too. Works pretty good. Even fixed a bug in antialias, so vectorblur will be better. Also: found out that OpenGL display list speedup accidentally was still triggered with the rt button... so it did not work by default.
2006-12-10 20:30:15 +00:00
case SCE_PASS_UV:
Bugfix #5712 UV rendering error: the third value of a UV texture coordinate should be set to zero always, this enables 3d textures to work uniform with UV. Also fixed: UV pass output for non_OSA is now equal to OSA, meaning only used UV face values are in such passes.
2007-01-15 11:41:10 +00:00
if(shi->totuv) {
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
uvcol[0]= 0.5f + 0.5f*shi->uv[shi->actuv].uv[0];
uvcol[1]= 0.5f + 0.5f*shi->uv[shi->actuv].uv[1];
Bugfix #5712 UV rendering error: the third value of a UV texture coordinate should be set to zero always, this enables 3d textures to work uniform with UV. Also fixed: UV pass output for non_OSA is now equal to OSA, meaning only used UV face values are in such passes.
2007-01-15 11:41:10 +00:00
uvcol[2]= 1.0f;
col= uvcol;
}
Composite & Pass render goodies: - New Passes: UV and Rad(iosity) - New Nodes: UV Map and Index Mask - Z-combine now is antialiased As usual, please check the log. Has nice pics! http://www.blender3d.org/cms/Composite__UV_Map__ID.830.0.html For devs: the antialias code from Vector Blur is now exported in compo too. Works pretty good. Even fixed a bug in antialias, so vectorblur will be better. Also: found out that OpenGL display list speedup accidentally was still triggered with the rt button... so it did not work by default.
2006-12-10 20:30:15 +00:00
break;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
case SCE_PASS_VECTOR:
col= shr->winspeed;
iImage based Vector Blur After a couple of experiments with variable blur filters, I tried a more interesting, and who knows... original approach. :) First watch results here: http://www.blender.org/bf/rt0001_0030.avi http://www.blender.org/bf/hand0001_0060.avi These are the steps in producing such results: - In preprocess, the speed vectors to previous and next frame are calculated. Speed vectors are screen-aligned and in pixel size. - while rendering, these vectors get calculated per sample, and accumulated in the vector buffer checking for "minimum speed". (on start the vector buffer is initialized on max speed). - After render: - The entire image, all pixels, then is converted to quad polygons. - Also the z value of the pixels is assigned to the polygons - The vertices for the quads use averaged speed vectors (of the 4 corner faces), using a 'minimum but non-zero' speed rule. This minimal speed trick works very well to prevent 'tearing' apart when multiple faces move in different directions in a pixel, or to be able to separate moving pixels clearly from non-moving ones - So, now we have a sort of 'mask' of quad polygons. The previous steps guaranteed that this mask doesn't have antialias color info, and has speed vectors that ensure individual parts to move nicely without tearing effects. The Z allows multiple layers of moving masks. - Then, in temporal buffer, faces get tagged if they move or not - These tags then go to an anti-alias routine, which assigns alpha values to edge faces, based on the method we used in past to antialias bitmaps (still in our code, check the antialias.c in imbuf!) - finally, the tag buffer is used to tag which z values of the original image have to be included (to allow blur go behind stuff). - OK, now we're ready for accumulating! In a loop, all faces then get drawn (with zbuffer) with increasing influence of their speed vectors. The resulting image then is accumulated on top of the original with a decreasing weighting value. It sounds all quite complex... but the speed is still encouraging. Above images have 64 mblur steps, which takes about 1-3 seconds per frame. Usage notes: - Make sure the render-layer has passes 'Vector' and 'Z' on. - add in Compositor the VectorBlur node, and connect the image, Z and speed to the inputs. - The node allows to set amount of steps (10 steps = 10 forward, 10 back). and to set a maximum speed in pixels... to prevent extreme moving things to blur too wide.
2006-02-06 22:11:50 +00:00
pixsize= 4;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
break;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
case SCE_PASS_INDEXOB:
if(shi->vlr) {
fp= rpass->rect + offset;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
*fp= (float)shi->obr->ob->index;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
}
break;
Two fixes for yesterday commit, new Mist pass: - non OSA case didn't work - ztransp adding was accidentally using an incorrect alpha value NOTE: allmost all pass types rendered in OSA with a filter (not box!) were incorrectly added on solid layers. Like diffuse, AO, etc.
2008-01-20 14:50:30 +00:00
case SCE_PASS_MIST:
fp= rpass->rect + offset;
*fp= shr->mist;
break;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
}
if(col) {
fp= rpass->rect + pixsize*offset;
for(a=0; a<pixsize; a++)
fp[a]= col[a];
}
}
}
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
int get_sample_layers(RenderPart *pa, RenderLayer *rl, RenderLayer **rlpp)
{
if(pa->fullresult.first) {
int sample, nr= BLI_findindex(&pa->result->layers, rl);
for(sample=0; sample<R.osa; sample++) {
RenderResult *rr= BLI_findlink(&pa->fullresult, sample);
rlpp[sample]= BLI_findlink(&rr->layers, nr);
}
return R.osa;
}
else {
rlpp[0]= rl;
return 1;
}
}
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
/* only do sky, is default in the solid layer (shade_tile) btw */
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
static void sky_tile(RenderPart *pa, RenderLayer *rl)
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
{
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
RenderLayer *rlpp[RE_MAX_OSA];
int x, y, od=0, totsample;
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
if(R.r.alphamode!=R_ADDSKY)
return;
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
totsample= get_sample_layers(pa, rl, rlpp);
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
for(y=pa->disprect.ymin; y<pa->disprect.ymax; y++) {
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
for(x=pa->disprect.xmin; x<pa->disprect.xmax; x++, od+=4) {
float col[4];
int sample, done= 0;
for(sample= 0; sample<totsample; sample++) {
float *pass= rlpp[sample]->rectf + od;
if(pass[3]<1.0f) {
if(done==0) {
2.5: merge with trunk, previous merge was only up to yesterday. svn merge https://svn.blender.org/svnroot/bf-blender/trunk/blender -r17416:HEAD
2008-11-12 22:03:11 +00:00
shadeSkyPixel(col, x, y, pa->thread);
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
done= 1;
}
if(pass[3]==0.0f) {
QUATCOPY(pass, col);
}
else {
addAlphaUnderFloat(pass, col);
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
}
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
}
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
if(y&1)
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(R.test_break(R.tbh)) break;
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
}
}
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
static void atm_tile(RenderPart *pa, RenderLayer *rl)
{
RenderPass *zpass;
GroupObject *go;
LampRen *lar;
Bug Fix #17337, now atmosphere works correctly with FSA enabled
2008-07-16 08:10:23 +00:00
RenderLayer *rlpp[RE_MAX_OSA];
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
int totsample;
int x, y, od= 0;
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
Bug Fix #17337, now atmosphere works correctly with FSA enabled
2008-07-16 08:10:23 +00:00
totsample= get_sample_layers(pa, rl, rlpp);
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
/* check that z pass is enabled */
if(pa->rectz==NULL) return;
for(zpass= rl->passes.first; zpass; zpass= zpass->next)
if(zpass->passtype==SCE_PASS_Z)
break;
if(zpass==NULL) return;
/* check for at least one sun lamp that its atmosphere flag is is enabled */
for(go=R.lights.first; go; go= go->next) {
lar= go->lampren;
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
if(lar->type==LA_SUN && lar->sunsky && (lar->sunsky->effect_type & LA_SUN_EFFECT_AP))
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
break;
}
/* do nothign and return if there is no sun lamp */
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
if(go==NULL)
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
return;
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
/* for each x,y and each sample, and each sun lamp*/
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
for(y=pa->disprect.ymin; y<pa->disprect.ymax; y++) {
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
for(x=pa->disprect.xmin; x<pa->disprect.xmax; x++, od++) {
int sample;
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
for(sample=0; sample<totsample; sample++) {
float *zrect= RE_RenderLayerGetPass(rlpp[sample], SCE_PASS_Z) + od;
float *rgbrect = rlpp[sample]->rectf + 4*od;
2.5 Material: * Added material "type" property, with Surface/Volume/Halo options, compatible with sim_physics, as requested for material buttons layout. Obviously the Volume setting does nothing currently. * Deprecated MA_HALO flag in favor of this.
2009-05-20 13:34:04 +00:00
float rgb[3] = {0};
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
int done= 0;
for(go=R.lights.first; go; go= go->next) {
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
lar= go->lampren;
if(lar->type==LA_SUN && lar->sunsky) {
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
/* if it's sky continue and don't apply atmosphere effect on it */
Bugfix #17656 - Changed order for applying atmosphere, it does it now before alpha-adding sky, giving correct transparency - Added correction for de-premulling and premulling scatter color
2008-09-21 12:03:41 +00:00
if(*zrect >= 9.9e10 || rgbrect[3]==0.0f) {
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
continue;
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
}
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
if((lar->sunsky->effect_type & LA_SUN_EFFECT_AP)) {
float tmp_rgb[3];
VECCOPY(tmp_rgb, rgbrect);
Bugfix #17656 - Changed order for applying atmosphere, it does it now before alpha-adding sky, giving correct transparency - Added correction for de-premulling and premulling scatter color
2008-09-21 12:03:41 +00:00
if(rgbrect[3]!=1.0f) { /* de-premul */
float div= 1.0f/rgbrect[3];
VECMUL(tmp_rgb, div);
}
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
shadeAtmPixel(lar->sunsky, tmp_rgb, x, y, *zrect);
Bugfix #17656 - Changed order for applying atmosphere, it does it now before alpha-adding sky, giving correct transparency - Added correction for de-premulling and premulling scatter color
2008-09-21 12:03:41 +00:00
if(rgbrect[3]!=1.0f) { /* premul */
VECMUL(tmp_rgb, rgbrect[3]);
}
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
if(done==0) {
VECCOPY(rgb, tmp_rgb);
done = 1;
}
else{
rgb[0] = 0.5f*rgb[0] + 0.5f*tmp_rgb[0];
rgb[1] = 0.5f*rgb[1] + 0.5f*tmp_rgb[1];
rgb[2] = 0.5f*rgb[2] + 0.5f*tmp_rgb[2];
}
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
}
}
}
#bugfix #17337 FSA didn't work for atmosphere option.
2008-09-21 10:31:22 +00:00
/* if at least for one sun lamp aerial perspective was applied*/
if(done) {
VECCOPY(rgbrect, rgb);
Bug Fix #17337, now atmosphere works correctly with FSA enabled
2008-07-16 08:10:23 +00:00
}
}
Sun,Sky and atmosphere for lamps(sun type), PATCH#8063 http://projects.blender.org/tracker/?func=detail&atid=127&aid=8063&group_id=9
2008-07-03 10:38:35 +00:00
}
}
}
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
static void shadeDA_tile(RenderPart *pa, RenderLayer *rl)
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
{
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
RenderResult *rr= pa->result;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
ShadeSample ssamp;
Win64: please check my changes if you ran across them ;) But should be fine since no additional crashes were reported!
2008-08-17 17:08:00 +00:00
intptr_t *rd, *rectdaps= pa->rectdaps;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
int samp;
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
int x, y, seed, crop=0, offs=0, od;
Random() issues with rendering... - AO and soft shadow AreaLight tables were generated without fixed seed, causing animations to give unwanted amounts of noise. - Made sure these tables now are calculated before render, with fixed seed - Then found out the BLI_rand() has very bad seeding... it showed up as patterns. After some experimenting, found a nice method using noise.c hash tables. For compatibility with old code, named it BLI_srandom() to use this next to the BLI_srand(). This follows libc rand() and random() naming convention. - Then of course threading should work... so made a BLI_thread_rand version of the calls. Now supports up to 16 threads, comments added in .h and .c Result is stable animation render with AO and soft shadow. But, please test and feedback!
2005-08-25 13:11:04 +00:00
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(R.test_break(R.tbh)) return;
Random() issues with rendering... - AO and soft shadow AreaLight tables were generated without fixed seed, causing animations to give unwanted amounts of noise. - Made sure these tables now are calculated before render, with fixed seed - Then found out the BLI_rand() has very bad seeding... it showed up as patterns. After some experimenting, found a nice method using noise.c hash tables. For compatibility with old code, named it BLI_srandom() to use this next to the BLI_srand(). This follows libc rand() and random() naming convention. - Then of course threading should work... so made a BLI_thread_rand version of the calls. Now supports up to 16 threads, comments added in .h and .c Result is stable animation render with AO and soft shadow. But, please test and feedback!
2005-08-25 13:11:04 +00:00
New shadow feature: Irregular Shadow Buffers Full log: http://www.blender3d.org/cms/Irregular_Shadow_Buffe.785.0.html In short: this is a shadow buffer approach that always results in crispy shadows, independent of lamp buffer size or zoom level. This shadow buffer system also supports transparent shadow. This is part of work on refreshing Shadow Buffers in Blender. You now can choose of two types (Classical, Irregular). More types will follow. Also quality issues for Classical shadow buffers are going to be reviewed, especially to solve the lousy Biasing. For the CVS log record; it is based on articles: Gregory Johnson et al, University of Texas, Austin. (Regular grid method). Timo Aila and Samuli Laine, Helsinki University of Technology. (BSP method).
2006-10-14 10:21:19 +00:00
/* irregular shadowb buffer creation */
if(R.r.mode & R_SHADOW)
ISB_create(pa, NULL);
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* we set per pixel a fixed seed, for random AO and shadow samples */
seed= pa->rectx*pa->disprect.ymin;
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* general shader info, passes */
shade_sample_initialize(&ssamp, pa, rl);
Approximate Ambient Occlusion ============================= A new approximate ambient occlusion method has been added, next to the existing one based on raytracing. This method is specifically targetted at use in animations, since it is inherently noise free, and so will not flicker across frames. http://www.blender.org/development/current-projects/changes-since-244/approximate-ambient-occlusion/ http://peach.blender.org/index.php/approximate-ambient-occlusion/ Further improvements are still needed, but it can be tested already. There are still a number of known issues: - Bias errors on backfaces. - For performance, instanced object do not occlude currently. - Sky textures don't work well, the derivatives for texture evaluation are not correct. - Multiple passes do not work entirely correct (they are not accurate to begin with, but could be better).
2008-01-17 19:27:16 +00:00
/* occlusion caching */
if(R.occlusiontree)
cache_occ_samples(&R, pa, &ssamp);
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* filtered render, for now we assume only 1 filter size */
if(pa->crop) {
crop= 1;
rectdaps+= pa->rectx + 1;
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
offs= pa->rectx + 1;
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
}
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
/* scanline updates have to be 2 lines behind */
rr->renrect.ymin= 0;
rr->renrect.ymax= -2*crop;
Carefully went over all scanline updating while rendering, to ensure only updates are allowed to draw when a part is within a scanline rendering loop. Might solve threads issues with opengl...
2006-02-21 10:47:08 +00:00
rr->renlay= rl;
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
for(y=pa->disprect.ymin+crop; y<pa->disprect.ymax-crop; y++, rr->renrect.ymax++) {
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
rd= rectdaps;
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
od= offs;
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
for(x=pa->disprect.xmin+crop; x<pa->disprect.xmax-crop; x++, rd++, od++) {
Bugfix, own collection: Random seeding is still not perfect in render, especially lack of good thread support still. - VectorBlur node was calling seed for each exec, causing other nodes to get fixed random too. - added seed in non-OSA main loop for render - use BLI_srandom, is better than BLI_srand
2006-11-21 15:52:45 +00:00
BLI_thread_srandom(pa->thread, seed++);
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
if(*rd) {
if(shade_samples(&ssamp, (PixStr *)(*rd), x, y)) {
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
/* multisample buffers or filtered mask filling? */
if(pa->fullresult.first) {
int a;
for(samp=0; samp<ssamp.tot; samp++) {
int smask= ssamp.shi[samp].mask;
for(a=0; a<R.osa; a++) {
int mask= 1<<a;
if(smask & mask)
add_passes(ssamp.rlpp[a], od, &ssamp.shi[samp], &ssamp.shr[samp]);
}
}
}
else {
for(samp=0; samp<ssamp.tot; samp++)
Composite & Pass render goodies: - New Passes: UV and Rad(iosity) - New Nodes: UV Map and Index Mask - Z-combine now is antialiased As usual, please check the log. Has nice pics! http://www.blender3d.org/cms/Composite__UV_Map__ID.830.0.html For devs: the antialias code from Vector Blur is now exported in compo too. Works pretty good. Even fixed a bug in antialias, so vectorblur will be better. Also: found out that OpenGL display list speedup accidentally was still triggered with the rt button... so it did not work by default.
2006-12-10 20:30:15 +00:00
add_filt_passes(rl, ssamp.shi[samp].mask, pa->rectx, od, &ssamp.shi[samp], &ssamp.shr[samp]);
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
}
}
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
}
}
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
rectdaps+= pa->rectx;
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
offs+= pa->rectx;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(y&1) if(R.test_break(R.tbh)) break;
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
}
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
Carefully went over all scanline updating while rendering, to ensure only updates are allowed to draw when a part is within a scanline rendering loop. Might solve threads issues with opengl...
2006-02-21 10:47:08 +00:00
/* disable scanline updating */
rr->renlay= NULL;
New shadow feature: Irregular Shadow Buffers Full log: http://www.blender3d.org/cms/Irregular_Shadow_Buffe.785.0.html In short: this is a shadow buffer approach that always results in crispy shadows, independent of lamp buffer size or zoom level. This shadow buffer system also supports transparent shadow. This is part of work on refreshing Shadow Buffers in Blender. You now can choose of two types (Classical, Irregular). More types will follow. Also quality issues for Classical shadow buffers are going to be reviewed, especially to solve the lousy Biasing. For the CVS log record; it is based on articles: Gregory Johnson et al, University of Texas, Austin. (Regular grid method). Timo Aila and Samuli Laine, Helsinki University of Technology. (BSP method).
2006-10-14 10:21:19 +00:00
if(R.r.mode & R_SHADOW)
ISB_free(pa);
Approximate Ambient Occlusion ============================= A new approximate ambient occlusion method has been added, next to the existing one based on raytracing. This method is specifically targetted at use in animations, since it is inherently noise free, and so will not flicker across frames. http://www.blender.org/development/current-projects/changes-since-244/approximate-ambient-occlusion/ http://peach.blender.org/index.php/approximate-ambient-occlusion/ Further improvements are still needed, but it can be tested already. There are still a number of known issues: - Bias errors on backfaces. - For performance, instanced object do not occlude currently. - Sky textures don't work well, the derivatives for texture evaluation are not correct. - Multiple passes do not work entirely correct (they are not accurate to begin with, but could be better).
2008-01-17 19:27:16 +00:00
if(R.occlusiontree)
free_occ_samples(&R, pa);
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* ************* pixel struct ******** */
Initial revision
2002-10-12 11:37:38 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
static PixStrMain *addpsmain(ListBase *lb)
{
PixStrMain *psm;
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
psm= (PixStrMain *)MEM_mallocN(sizeof(PixStrMain),"pixstrMain");
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
BLI_addtail(lb, psm);
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
psm->ps= (PixStr *)MEM_mallocN(4096*sizeof(PixStr),"pixstr");
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
psm->counter= 0;
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
return psm;
}
Initial revision
2002-10-12 11:37:38 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
static void freeps(ListBase *lb)
{
PixStrMain *psm, *psmnext;
for(psm= lb->first; psm; psm= psmnext) {
psmnext= psm->next;
if(psm->ps)
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
MEM_freeN(psm->ps);
MEM_freeN(psm);
Initial revision
2002-10-12 11:37:38 +00:00
}
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
lb->first= lb->last= NULL;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
Win64: please check my changes if you ran across them ;) But should be fine since no additional crashes were reported!
2008-08-17 17:08:00 +00:00
static void addps(ListBase *lb, intptr_t *rd, int obi, int facenr, int z, int maskz, unsigned short mask)
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
{
PixStrMain *psm;
PixStr *ps, *last= NULL;
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
if(*rd) {
ps= (PixStr *)(*rd);
while(ps) {
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
if( ps->obi == obi && ps->facenr == facenr ) {
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
ps->mask |= mask;
return;
Initial revision
2002-10-12 11:37:38 +00:00
}
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
last= ps;
ps= ps->next;
Initial revision
2002-10-12 11:37:38 +00:00
}
}
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* make new PS (pixel struct) */
psm= lb->last;
if(psm->counter==4095)
psm= addpsmain(lb);
ps= psm->ps + psm->counter++;
if(last) last->next= ps;
Win64: please check my changes if you ran across them ;) But should be fine since no additional crashes were reported!
2008-08-17 17:08:00 +00:00
else *rd= (intptr_t)ps;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
ps->next= NULL;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
ps->obi= obi;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
ps->facenr= facenr;
ps->z= z;
Negative zmask rendering: now there is a setting to render everything behind the zmask instead of in front. Might need tweaks, and doesn't work with halo yet.
2008-02-20 14:17:53 +00:00
ps->maskz= maskz;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
ps->mask = mask;
Uninitialized variable fix. Might fix bug #6431, error in rendering of irregular shadowbuffers. Found by valgrind session, provided by Mathias 'Lynx3d' Wein
2007-03-26 19:44:21 +00:00
ps->shadfac= 0;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
static void edge_enhance_add(RenderPart *pa, float *rectf, float *arect)
{
float addcol[4];
int pix;
if(arect==NULL)
return;
for(pix= pa->rectx*pa->recty; pix>0; pix--, arect++, rectf+=4) {
if(*arect != 0.0f) {
addcol[0]= *arect * R.r.edgeR;
addcol[1]= *arect * R.r.edgeG;
addcol[2]= *arect * R.r.edgeB;
addcol[3]= *arect;
addAlphaOverFloat(rectf, addcol);
}
}
}
Bugfix #17497 FSA didn't support "key alpha", fixed it. Also noticed edge-enhance didnt work for FSA, added warning for that, fixing it is too clumsy, this option belongs in compositor.
2008-09-07 13:15:44 +00:00
static void convert_to_key_alpha(RenderPart *pa, RenderLayer *rl)
Bugfix, IRC report: Render option "Key Alpha" did not work yet.
2006-06-29 17:22:29 +00:00
{
Bugfix #17497 FSA didn't support "key alpha", fixed it. Also noticed edge-enhance didnt work for FSA, added warning for that, fixing it is too clumsy, this option belongs in compositor.
2008-09-07 13:15:44 +00:00
RenderLayer *rlpp[RE_MAX_OSA];
int y, sample, totsample;
totsample= get_sample_layers(pa, rl, rlpp);
Bugfix, IRC report: Render option "Key Alpha" did not work yet.
2006-06-29 17:22:29 +00:00
Bugfix #17497 FSA didn't support "key alpha", fixed it. Also noticed edge-enhance didnt work for FSA, added warning for that, fixing it is too clumsy, this option belongs in compositor.
2008-09-07 13:15:44 +00:00
for(sample= 0; sample<totsample; sample++) {
float *rectf= rlpp[sample]->rectf;
for(y= pa->rectx*pa->recty; y>0; y--, rectf+=4) {
if(rectf[3] >= 1.0f);
else if(rectf[3] > 0.0f) {
rectf[0] /= rectf[3];
rectf[1] /= rectf[3];
rectf[2] /= rectf[3];
}
Bugfix, IRC report: Render option "Key Alpha" did not work yet.
2006-06-29 17:22:29 +00:00
}
}
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* adds only alpha values */
Fix for bug #10990: edge render AA issue.
2008-05-07 14:58:33 +00:00
void edge_enhance_tile(RenderPart *pa, float *rectf, int *rectz)
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
{
/* use zbuffer to define edges, add it to the image */
int y, x, col, *rz, *rz1, *rz2, *rz3;
int zval1, zval2, zval3;
float *rf;
/* shift values in zbuffer 4 to the right (anti overflows), for filter we need multiplying with 12 max */
Fix for bug #10990: edge render AA issue.
2008-05-07 14:58:33 +00:00
rz= rectz;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
if(rz==NULL) return;
for(y=0; y<pa->recty; y++)
for(x=0; x<pa->rectx; x++, rz++) (*rz)>>= 4;
Fix for bug #10990: edge render AA issue.
2008-05-07 14:58:33 +00:00
rz1= rectz;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
rz2= rz1+pa->rectx;
rz3= rz2+pa->rectx;
rf= rectf+pa->rectx+1;
for(y=0; y<pa->recty-2; y++) {
for(x=0; x<pa->rectx-2; x++, rz1++, rz2++, rz3++, rf++) {
/* prevent overflow with sky z values */
zval1= rz1[0] + 2*rz1[1] + rz1[2];
zval2= 2*rz2[0] + 2*rz2[2];
zval3= rz3[0] + 2*rz3[1] + rz3[2];
col= ( 4*rz2[1] - (zval1 + zval2 + zval3)/3 );
if(col<0) col= -col;
col >>= 5;
if(col > (1<<16)) col= (1<<16);
else col= (R.r.edgeint*col)>>8;
if(col>0) {
float fcol;
if(col>255) fcol= 1.0f;
else fcol= (float)col/255.0f;
if(R.osa)
*rf+= fcol/(float)R.osa;
else
*rf= fcol;
}
}
rz1+= 2;
rz2+= 2;
rz3+= 2;
rf+= 2;
}
/* shift back zbuf values, we might need it still */
Fix for bug #10990: edge render AA issue.
2008-05-07 14:58:33 +00:00
rz= rectz;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
for(y=0; y<pa->recty; y++)
for(x=0; x<pa->rectx; x++, rz++) (*rz)<<= 4;
}
bugfix #5755 Render: Vector blur didn't work proper for Ztransp (accidentally was antialising speed, which should not happen). Also added speed-pass clearing in render, this to elimate the 'tsk tsk' print in the Vector Blur node. :)
2007-01-21 10:04:50 +00:00
static void reset_sky_speed(RenderPart *pa, RenderLayer *rl)
{
/* for all pixels with max speed, set to zero */
Bugfix: FSA enabled in compositing scene but disabled in other scenes could crash, there was code to make sure osa level is the same in all scenes, but that was set too late, after sample tables are created. Fix for some unitinialized vector warnings with FSA, these were harmless, unfortunately.
2008-02-11 11:00:09 +00:00
RenderLayer *rlpp[RE_MAX_OSA];
bugfix #5755 Render: Vector blur didn't work proper for Ztransp (accidentally was antialising speed, which should not happen). Also added speed-pass clearing in render, this to elimate the 'tsk tsk' print in the Vector Blur node. :)
2007-01-21 10:04:50 +00:00
float *fp;
Bugfix: FSA enabled in compositing scene but disabled in other scenes could crash, there was code to make sure osa level is the same in all scenes, but that was set too late, after sample tables are created. Fix for some unitinialized vector warnings with FSA, these were harmless, unfortunately.
2008-02-11 11:00:09 +00:00
int a, sample, totsample;
bugfix #5755 Render: Vector blur didn't work proper for Ztransp (accidentally was antialising speed, which should not happen). Also added speed-pass clearing in render, this to elimate the 'tsk tsk' print in the Vector Blur node. :)
2007-01-21 10:04:50 +00:00
Bugfix: FSA enabled in compositing scene but disabled in other scenes could crash, there was code to make sure osa level is the same in all scenes, but that was set too late, after sample tables are created. Fix for some unitinialized vector warnings with FSA, these were harmless, unfortunately.
2008-02-11 11:00:09 +00:00
totsample= get_sample_layers(pa, rl, rlpp);
bugfix #5755 Render: Vector blur didn't work proper for Ztransp (accidentally was antialising speed, which should not happen). Also added speed-pass clearing in render, this to elimate the 'tsk tsk' print in the Vector Blur node. :)
2007-01-21 10:04:50 +00:00
Bugfix: FSA enabled in compositing scene but disabled in other scenes could crash, there was code to make sure osa level is the same in all scenes, but that was set too late, after sample tables are created. Fix for some unitinialized vector warnings with FSA, these were harmless, unfortunately.
2008-02-11 11:00:09 +00:00
for(sample= 0; sample<totsample; sample++) {
fp= RE_RenderLayerGetPass(rlpp[sample], SCE_PASS_VECTOR);
if(fp==NULL) break;
for(a= 4*pa->rectx*pa->recty - 1; a>=0; a--)
if(fp[a] == PASS_VECTOR_MAX) fp[a]= 0.0f;
}
}
bugfix #5755 Render: Vector blur didn't work proper for Ztransp (accidentally was antialising speed, which should not happen). Also added speed-pass clearing in render, this to elimate the 'tsk tsk' print in the Vector Blur node. :)
2007-01-21 10:04:50 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
static unsigned short *make_solid_mask(RenderPart *pa)
{
Win64: please check my changes if you ran across them ;) But should be fine since no additional crashes were reported!
2008-08-17 17:08:00 +00:00
intptr_t *rd= pa->rectdaps;
Fix numerous gcc warnings.
2007-03-16 05:03:38 +00:00
unsigned short *solidmask, *sp;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
int x;
if(rd==NULL) return NULL;
sp=solidmask= MEM_mallocN(sizeof(short)*pa->rectx*pa->recty, "solidmask");
for(x=pa->rectx*pa->recty; x>0; x--, rd++, sp++) {
if(*rd) {
PixStr *ps= (PixStr *)*rd;
*sp= ps->mask;
for(ps= ps->next; ps; ps= ps->next)
*sp |= ps->mask;
}
else
*sp= 0;
}
return solidmask;
}
static void addAlphaOverFloatMask(float *dest, float *source, unsigned short dmask, unsigned short smask)
{
unsigned short shared= dmask & smask;
float mul= 1.0 - source[3];
if(shared) { /* overlapping masks */
/* masks differ, we make a mixture of 'add' and 'over' */
if(shared!=dmask) {
float shared_bits= (float)count_mask(shared); /* alpha over */
float tot_bits= (float)count_mask(smask|dmask); /* alpha add */
float add= (tot_bits - shared_bits)/tot_bits; /* add level */
mul= add + (1.0f-add)*mul;
}
}
else if(dmask && smask) {
/* works for premul only, of course */
dest[0]+= source[0];
dest[1]+= source[1];
dest[2]+= source[2];
dest[3]+= source[3];
return;
}
dest[0]= (mul*dest[0]) + source[0];
dest[1]= (mul*dest[1]) + source[1];
dest[2]= (mul*dest[2]) + source[2];
dest[3]= (mul*dest[3]) + source[3];
}
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
typedef struct ZbufSolidData {
RenderLayer *rl;
ListBase *psmlist;
float *edgerect;
} ZbufSolidData;
void make_pixelstructs(RenderPart *pa, ZSpan *zspan, int sample, void *data)
{
ZbufSolidData *sdata= (ZbufSolidData*)data;
ListBase *lb= sdata->psmlist;
Win64: please check my changes if you ran across them ;) But should be fine since no additional crashes were reported!
2008-08-17 17:08:00 +00:00
intptr_t *rd= pa->rectdaps;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
int *ro= zspan->recto;
int *rp= zspan->rectp;
int *rz= zspan->rectz;
Negative zmask rendering: now there is a setting to render everything behind the zmask instead of in front. Might need tweaks, and doesn't work with halo yet.
2008-02-20 14:17:53 +00:00
int *rm= zspan->rectmask;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
int x, y;
int mask= 1<<sample;
for(y=0; y<pa->recty; y++) {
Negative zmask rendering: now there is a setting to render everything behind the zmask instead of in front. Might need tweaks, and doesn't work with halo yet.
2008-02-20 14:17:53 +00:00
for(x=0; x<pa->rectx; x++, rd++, rp++, ro++, rz++, rm++) {
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
if(*rp) {
Negative zmask rendering: now there is a setting to render everything behind the zmask instead of in front. Might need tweaks, and doesn't work with halo yet.
2008-02-20 14:17:53 +00:00
addps(lb, rd, *ro, *rp, *rz, (zspan->rectmask)? *rm: 0, mask);
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
}
}
}
if(sdata->rl->layflag & SCE_LAY_EDGE)
if(R.r.mode & R_EDGE)
Fix for bug #10990: edge render AA issue.
2008-05-07 14:58:33 +00:00
edge_enhance_tile(pa, sdata->edgerect, zspan->rectz);
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* main call for shading Delta Accum, for OSA */
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* supposed to be fully threadable! */
void zbufshadeDA_tile(RenderPart *pa)
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
{
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
RenderResult *rr= pa->result;
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
RenderLayer *rl;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
ListBase psmlist= {NULL, NULL};
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
float *edgerect= NULL;
Random() issues with rendering... - AO and soft shadow AreaLight tables were generated without fixed seed, causing animations to give unwanted amounts of noise. - Made sure these tables now are calculated before render, with fixed seed - Then found out the BLI_rand() has very bad seeding... it showed up as patterns. After some experimenting, found a nice method using noise.c hash tables. For compatibility with old code, named it BLI_srandom() to use this next to the BLI_srand(). This follows libc rand() and random() naming convention. - Then of course threading should work... so made a BLI_thread_rand version of the calls. Now supports up to 16 threads, comments added in .h and .c Result is stable animation render with AO and soft shadow. But, please test and feedback!
2005-08-25 13:11:04 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* allocate the necessary buffers */
/* zbuffer inits these rects */
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
pa->recto= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "recto");
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
pa->rectp= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "rectp");
pa->rectz= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "rectz");
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
for(rl= rr->layers.first; rl; rl= rl->next) {
Negative zmask rendering: now there is a setting to render everything behind the zmask instead of in front. Might need tweaks, and doesn't work with halo yet.
2008-02-20 14:17:53 +00:00
if((rl->layflag & SCE_LAY_ZMASK) && (rl->layflag & SCE_LAY_NEG_ZMASK))
pa->rectmask= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "rectmask");
Bug 4429 When using multiple RenderLayers, each with own 'edge render', the buffer holding edge pixels wasn't cleared from scratch, so the edges of previous layers were showing too.
2006-06-23 16:48:28 +00:00
/* initialize pixelstructs and edge buffer */
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
addpsmain(&psmlist);
Win64: please check my changes if you ran across them ;) But should be fine since no additional crashes were reported!
2008-08-17 17:08:00 +00:00
pa->rectdaps= MEM_callocN(sizeof(intptr_t)*pa->rectx*pa->recty+4, "zbufDArectd");
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
if(rl->layflag & SCE_LAY_EDGE)
if(R.r.mode & R_EDGE)
edgerect= MEM_callocN(sizeof(float)*pa->rectx*pa->recty, "rectedge");
Random() issues with rendering... - AO and soft shadow AreaLight tables were generated without fixed seed, causing animations to give unwanted amounts of noise. - Made sure these tables now are calculated before render, with fixed seed - Then found out the BLI_rand() has very bad seeding... it showed up as patterns. After some experimenting, found a nice method using noise.c hash tables. For compatibility with old code, named it BLI_srandom() to use this next to the BLI_srand(). This follows libc rand() and random() naming convention. - Then of course threading should work... so made a BLI_thread_rand version of the calls. Now supports up to 16 threads, comments added in .h and .c Result is stable animation render with AO and soft shadow. But, please test and feedback!
2005-08-25 13:11:04 +00:00
Tweaks in RenderLayer rendering: - LampHalos can be rendered separately too. Just disable 'Solid' in a layer and keep 'Halo' option enabled. - Note that disabling 'Solid' will still fill in Z values for the solid faces, to provide occlusion information for the Ztransp and Halo layer options. The latter didn't work this way until now for OSA render. ALso note that that Ztransp+LampHalo still isn't good marriage... it renders a bit weird, but that's an old issue. :)
2006-02-09 22:33:49 +00:00
/* always fill visibility */
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
for(pa->sample=0; pa->sample<R.osa; pa->sample+=4) {
ZbufSolidData sdata;
sdata.rl= rl;
sdata.psmlist= &psmlist;
sdata.edgerect= edgerect;
Phase one of better masking support while rendering. Problem: artist wants character to walk in grass, but still have all rendered in seperate render-layers, for postpro effects and vblur. How to efficiently create a mask image you can put *over* the character for the grass? Solution has two parts; this commits allows any layer inside of the renderlayers to become a Z-mask (Z values for solid gets filled in, but not rendered). Second part of commit is render option "Only render stuff that's in front of a zbuffer value that was filled in (saves render time)
2008-01-24 15:03:34 +00:00
zbuffer_solid(pa, rl, make_pixelstructs, &sdata);
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(R.test_break(R.tbh)) break;
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
}
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
/* shades solid */
if(rl->layflag & SCE_LAY_SOLID)
shadeDA_tile(pa, rl);
Tweaks in RenderLayer rendering: - LampHalos can be rendered separately too. Just disable 'Solid' in a layer and keep 'Halo' option enabled. - Note that disabling 'Solid' will still fill in Z values for the solid faces, to provide occlusion information for the Ztransp and Halo layer options. The latter didn't work this way until now for OSA render. ALso note that that Ztransp+LampHalo still isn't good marriage... it renders a bit weird, but that's an old issue. :)
2006-02-09 22:33:49 +00:00
/* lamphalo after solid, before ztra, looks nicest because ztra does own halo */
if(R.flag & R_LAMPHALO)
if(rl->layflag & SCE_LAY_HALO)
Numerous fixes in Render code: - Bug: material emit was ignored (showed in preview render backdrop) - Bug: world exposure was ignored - Bug: lamp halo was ignoring 'render layer light override'. Further reshuffled the way shadows are being pre-calculated, this to enable more advanced (and faster) usage of Material lightgroups. Now shadows are being cached in lamps, using a per-sample counter to check if a recalc is needed. Will also work (later) for Raytracing node shaders. - New: Material LightGroup option "Always", which always shades the lights in the group, independent of visibility layer. (so it allows to move such lights to hidden layer, not influencing anything).
2006-12-08 09:40:44 +00:00
lamphalo_tile(pa, rl);
Tweaks in RenderLayer rendering: - LampHalos can be rendered separately too. Just disable 'Solid' in a layer and keep 'Halo' option enabled. - Note that disabling 'Solid' will still fill in Z values for the solid faces, to provide occlusion information for the Ztransp and Halo layer options. The latter didn't work this way until now for OSA render. ALso note that that Ztransp+LampHalo still isn't good marriage... it renders a bit weird, but that's an old issue. :)
2006-02-09 22:33:49 +00:00
Fix #4087 Flipped layer render order for halo and ztransp. This because the ztransp layer changes z values now (for vblur). This means the default order is: solid + lamphalo + halo + ztra. Different combinations can be achieved with composting nodes.
2006-05-14 09:02:50 +00:00
/* halo before ztra, because ztra fills in zbuffer now */
if(R.flag & R_HALO)
if(rl->layflag & SCE_LAY_HALO)
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
halo_tile(pa, rl);
Particles ========= - Fix crash in particle transform with the particle system not editable. - Particle child distribution and caching is now multithreaded. - Child particles now have a separate Render Amount next to the existing Amount. The render amount particles are now only distributed and cached at render time, which should make editing with child particles faster. - Two new options for diffuse strand shading: - Surface Diffuse: computes the strand normal taking the normal at the surface into account. - Blending Distance: the distance in Blender units over which to blend in the normal at the surface. - Special strand rendering for more memory efficient and faster hair and grass. This is a work in progress, and has a number of known issues, don't report bugs to me for this feature yet. More info: http://www.blender.org/development/current-projects/changes-since-244/particles/
2007-12-04 13:57:28 +00:00
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
/* transp layer */
Strands now mix together correctly with ZTransp. They now also store a list of samples per pixel, and then get shaded together with the ztransp samples. This comes with a slight speed hit, but mainly memory might be a concern. However, testing some peach scenes I haven't problems.
2008-01-28 16:54:52 +00:00
if(R.flag & R_ZTRA || R.totstrand) {
if(rl->layflag & (SCE_LAY_ZTRA|SCE_LAY_STRAND)) {
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
if(pa->fullresult.first) {
zbuffer_transp_shade(pa, rl, rl->rectf, &psmlist);
}
else {
unsigned short *ztramask, *solidmask= NULL; /* 16 bits, MAX_OSA */
Bugfix #5491 Ztransp Mask issue: now I did all this work to solve the exceptions, but forgot the most simple case: regular alpha-over :) This solves for example AA issues in ztransp-hair over solid faces.
2006-12-22 12:43:40 +00:00
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
/* allocate, but not free here, for asynchronous display of this rect in main thread */
rl->acolrect= MEM_callocN(4*sizeof(float)*pa->rectx*pa->recty, "alpha layer");
/* swap for live updates, and it is used in zbuf.c!!! */
SWAP(float *, rl->acolrect, rl->rectf);
ztramask= zbuffer_transp_shade(pa, rl, rl->rectf, &psmlist);
SWAP(float *, rl->acolrect, rl->rectf);
/* zbuffer transp only returns ztramask if there's solid rendered */
if(ztramask)
solidmask= make_solid_mask(pa);
if(ztramask && solidmask) {
unsigned short *sps= solidmask, *spz= ztramask;
unsigned short fullmask= (1<<R.osa)-1;
float *fcol= rl->rectf; float *acol= rl->acolrect;
int x;
for(x=pa->rectx*pa->recty; x>0; x--, acol+=4, fcol+=4, sps++, spz++) {
if(*sps == fullmask)
addAlphaOverFloat(fcol, acol);
else
addAlphaOverFloatMask(fcol, acol, *sps, *spz);
}
Bugfix #5491 Ztransp Mask issue: now I did all this work to solve the exceptions, but forgot the most simple case: regular alpha-over :) This solves for example AA issues in ztransp-hair over solid faces.
2006-12-22 12:43:40 +00:00
}
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
else {
float *fcol= rl->rectf; float *acol= rl->acolrect;
int x;
for(x=pa->rectx*pa->recty; x>0; x--, acol+=4, fcol+=4) {
addAlphaOverFloat(fcol, acol);
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
}
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
if(solidmask) MEM_freeN(solidmask);
if(ztramask) MEM_freeN(ztramask);
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
}
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
}
}
Particles ========= - Fix crash in particle transform with the particle system not editable. - Particle child distribution and caching is now multithreaded. - Child particles now have a separate Render Amount next to the existing Amount. The render amount particles are now only distributed and cached at render time, which should make editing with child particles faster. - Two new options for diffuse strand shading: - Surface Diffuse: computes the strand normal taking the normal at the surface into account. - Blending Distance: the distance in Blender units over which to blend in the normal at the surface. - Special strand rendering for more memory efficient and faster hair and grass. This is a work in progress, and has a number of known issues, don't report bugs to me for this feature yet. More info: http://www.blender.org/development/current-projects/changes-since-244/particles/
2007-12-04 13:57:28 +00:00
Bugfix #17656 - Changed order for applying atmosphere, it does it now before alpha-adding sky, giving correct transparency - Added correction for de-premulling and premulling scatter color
2008-09-21 12:03:41 +00:00
/* sun/sky */
if(rl->layflag & SCE_LAY_SKY)
atm_tile(pa, rl);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* sky before edge */
if(rl->layflag & SCE_LAY_SKY)
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
sky_tile(pa, rl);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
/* extra layers */
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
if(rl->layflag & SCE_LAY_EDGE)
if(R.r.mode & R_EDGE)
edge_enhance_add(pa, rl->rectf, edgerect);
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
bugfix #5755 Render: Vector blur didn't work proper for Ztransp (accidentally was antialising speed, which should not happen). Also added speed-pass clearing in render, this to elimate the 'tsk tsk' print in the Vector Blur node. :)
2007-01-21 10:04:50 +00:00
if(rl->passflag & SCE_PASS_VECTOR)
reset_sky_speed(pa, rl);
Bugfix, IRC report: Render option "Key Alpha" did not work yet.
2006-06-29 17:22:29 +00:00
/* de-premul alpha */
if(R.r.alphamode & R_ALPHAKEY)
Bugfix #17497 FSA didn't support "key alpha", fixed it. Also noticed edge-enhance didnt work for FSA, added warning for that, fixing it is too clumsy, this option belongs in compositor.
2008-09-07 13:15:44 +00:00
convert_to_key_alpha(pa, rl);
Bugfix, IRC report: Render option "Key Alpha" did not work yet.
2006-06-29 17:22:29 +00:00
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
/* free stuff within loop! */
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
MEM_freeN(pa->rectdaps); pa->rectdaps= NULL;
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
freeps(&psmlist);
Bug 4429 When using multiple RenderLayers, each with own 'edge render', the buffer holding edge pixels wasn't cleared from scratch, so the edges of previous layers were showing too.
2006-06-23 16:48:28 +00:00
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
if(edgerect) MEM_freeN(edgerect);
Bug 4429 When using multiple RenderLayers, each with own 'edge render', the buffer holding edge pixels wasn't cleared from scratch, so the edges of previous layers were showing too.
2006-06-23 16:48:28 +00:00
edgerect= NULL;
Negative zmask rendering: now there is a setting to render everything behind the zmask instead of in front. Might need tweaks, and doesn't work with halo yet.
2008-02-20 14:17:53 +00:00
if(pa->rectmask) {
MEM_freeN(pa->rectmask);
pa->rectmask= NULL;
}
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
}
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* free all */
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
MEM_freeN(pa->recto); pa->recto= NULL;
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
MEM_freeN(pa->rectp); pa->rectp= NULL;
MEM_freeN(pa->rectz); pa->rectz= NULL;
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
Three feats; - RenderLayers with 'view layers' set, now also take visible lights into account. Works just like for scene layer settings. - On ESC from render, compositing (if set) is being skipped too - While rendering with multiple RenderLayers it will end with a display of the current RenderLayer (as in Scene buttons)
2006-02-02 21:08:39 +00:00
/* display active layer */
Carefully went over all scanline updating while rendering, to ensure only updates are allowed to draw when a part is within a scanline rendering loop. Might solve threads issues with opengl...
2006-02-21 10:47:08 +00:00
rr->renrect.ymin=rr->renrect.ymax= 0;
Rendering tricks for improving workflow: - Button option "Single" in render-layer panel will enable to only render the currently indicated render-layer. It will also skip compositing. - Brought back the 'Local View' render. This will only render the visible objects, but with lights from the original view-layers. To make the option useful, it also temporal enables 'Single', which has the a disadvantage that you need to set the correct render-layer. It is a bit a tricky option though... since its quite invisble and confusing for people who don't know the feature. This might become either a button in 3d header, or use a popup requester to confirm, or... will need to think over! At least; both options display in render window a text to denote the option.
2006-02-16 12:27:46 +00:00
rr->renlay= render_get_active_layer(&R, rr);
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
}
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* ------------------------------------------------------------------------ */
Initial revision
2002-10-12 11:37:38 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* non OSA case, full tile render */
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* supposed to be fully threadable! */
void zbufshade_tile(RenderPart *pa)
{
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
ShadeSample ssamp;
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
RenderResult *rr= pa->result;
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
RenderLayer *rl;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
PixStr ps;
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
float *edgerect= NULL;
features & fixes: - Enabled Groups to execute in Compositor. They were ignored still. Note; inside of groups nothing is cached, so a change of a group input will recalculate it fully. This is needed because groups are linked data (instances use same internal nodes). - Made Composit node "Viewer" display correctly input for images with 1/2/3/4 channels. - Added pass rendering, tested now with only regular Materials. For Material nodes this is quite more complex... since they cannot be easily separated in passes (each Material does a full shade) In this commit all pass render is disabled though, will continue work on that later. Sneak preview: http://www.blender.org/bf/rt.jpg (temporal image) - What did remain is the 'Normal' pass output. Normal works very nice for relighting effects. Use the "Normal Node" to define where more or less light should be. (Use "Value Map" node to tweak influence of the Normal node 'dot' output.) - EVIL bug fix: I've spend almost a day finding it... when combining AO and mirror render, the event queue was totally screwing up... two things not related at all! Found out error was in ray-mirror code, which was using partially uninitialized 'ShadeInput' data to pass on to render code. - Another fix; made sure that while thread render, the threads don't get events, only the main program will do. Might fix issues reported by people on linux/windows.
2006-02-02 17:54:22 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* fake pixel struct, to comply to osa render */
ps.next= NULL;
ps.mask= 0xFFFF;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
/* zbuffer code clears/inits rects */
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
pa->recto= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "recto");
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
pa->rectp= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "rectp");
pa->rectz= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "rectz");
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
for(rl= rr->layers.first; rl; rl= rl->next) {
Negative zmask rendering: now there is a setting to render everything behind the zmask instead of in front. Might need tweaks, and doesn't work with halo yet.
2008-02-20 14:17:53 +00:00
if((rl->layflag & SCE_LAY_ZMASK) && (rl->layflag & SCE_LAY_NEG_ZMASK))
pa->rectmask= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "rectmask");
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* general shader info, passes */
shade_sample_initialize(&ssamp, pa, rl);
Three feats; - RenderLayers with 'view layers' set, now also take visible lights into account. Works just like for scene layer settings. - On ESC from render, compositing (if set) is being skipped too - While rendering with multiple RenderLayers it will end with a display of the current RenderLayer (as in Scene buttons)
2006-02-02 21:08:39 +00:00
Phase one of better masking support while rendering. Problem: artist wants character to walk in grass, but still have all rendered in seperate render-layers, for postpro effects and vblur. How to efficiently create a mask image you can put *over* the character for the grass? Solution has two parts; this commits allows any layer inside of the renderlayers to become a Z-mask (Z values for solid gets filled in, but not rendered). Second part of commit is render option "Only render stuff that's in front of a zbuffer value that was filled in (saves render time)
2008-01-24 15:03:34 +00:00
zbuffer_solid(pa, rl, NULL, NULL);
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(!R.test_break(R.tbh)) { /* NOTE: this if() is not consistant */
New shadow feature: Irregular Shadow Buffers Full log: http://www.blender3d.org/cms/Irregular_Shadow_Buffe.785.0.html In short: this is a shadow buffer approach that always results in crispy shadows, independent of lamp buffer size or zoom level. This shadow buffer system also supports transparent shadow. This is part of work on refreshing Shadow Buffers in Blender. You now can choose of two types (Classical, Irregular). More types will follow. Also quality issues for Classical shadow buffers are going to be reviewed, especially to solve the lousy Biasing. For the CVS log record; it is based on articles: Gregory Johnson et al, University of Texas, Austin. (Regular grid method). Timo Aila and Samuli Laine, Helsinki University of Technology. (BSP method).
2006-10-14 10:21:19 +00:00
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
/* edges only for solid part, ztransp doesn't support it yet anti-aliased */
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
if(rl->layflag & SCE_LAY_EDGE) {
if(R.r.mode & R_EDGE) {
edgerect= MEM_callocN(sizeof(float)*pa->rectx*pa->recty, "rectedge");
Fix for bug #10990: edge render AA issue.
2008-05-07 14:58:33 +00:00
edge_enhance_tile(pa, edgerect, pa->rectz);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
}
}
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
/* initialize scanline updates for main thread */
rr->renrect.ymin= 0;
rr->renlay= rl;
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
if(rl->layflag & SCE_LAY_SOLID) {
float *fcol= rl->rectf;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
int *ro= pa->recto, *rp= pa->rectp, *rz= pa->rectz;
int x, y, offs=0, seed;
Bugfix, own collection: Random seeding is still not perfect in render, especially lack of good thread support still. - VectorBlur node was calling seed for each exec, causing other nodes to get fixed random too. - added seed in non-OSA main loop for render - use BLI_srandom, is better than BLI_srand
2006-11-21 15:52:45 +00:00
/* we set per pixel a fixed seed, for random AO and shadow samples */
seed= pa->rectx*pa->disprect.ymin;
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
New shadow feature: Irregular Shadow Buffers Full log: http://www.blender3d.org/cms/Irregular_Shadow_Buffe.785.0.html In short: this is a shadow buffer approach that always results in crispy shadows, independent of lamp buffer size or zoom level. This shadow buffer system also supports transparent shadow. This is part of work on refreshing Shadow Buffers in Blender. You now can choose of two types (Classical, Irregular). More types will follow. Also quality issues for Classical shadow buffers are going to be reviewed, especially to solve the lousy Biasing. For the CVS log record; it is based on articles: Gregory Johnson et al, University of Texas, Austin. (Regular grid method). Timo Aila and Samuli Laine, Helsinki University of Technology. (BSP method).
2006-10-14 10:21:19 +00:00
/* irregular shadowb buffer creation */
if(R.r.mode & R_SHADOW)
ISB_create(pa, NULL);
Approximate Ambient Occlusion ============================= A new approximate ambient occlusion method has been added, next to the existing one based on raytracing. This method is specifically targetted at use in animations, since it is inherently noise free, and so will not flicker across frames. http://www.blender.org/development/current-projects/changes-since-244/approximate-ambient-occlusion/ http://peach.blender.org/index.php/approximate-ambient-occlusion/ Further improvements are still needed, but it can be tested already. There are still a number of known issues: - Bias errors on backfaces. - For performance, instanced object do not occlude currently. - Sky textures don't work well, the derivatives for texture evaluation are not correct. - Multiple passes do not work entirely correct (they are not accurate to begin with, but could be better).
2008-01-17 19:27:16 +00:00
if(R.occlusiontree)
cache_occ_samples(&R, pa, &ssamp);
New shadow feature: Irregular Shadow Buffers Full log: http://www.blender3d.org/cms/Irregular_Shadow_Buffe.785.0.html In short: this is a shadow buffer approach that always results in crispy shadows, independent of lamp buffer size or zoom level. This shadow buffer system also supports transparent shadow. This is part of work on refreshing Shadow Buffers in Blender. You now can choose of two types (Classical, Irregular). More types will follow. Also quality issues for Classical shadow buffers are going to be reviewed, especially to solve the lousy Biasing. For the CVS log record; it is based on articles: Gregory Johnson et al, University of Texas, Austin. (Regular grid method). Timo Aila and Samuli Laine, Helsinki University of Technology. (BSP method).
2006-10-14 10:21:19 +00:00
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
for(y=pa->disprect.ymin; y<pa->disprect.ymax; y++, rr->renrect.ymax++) {
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
for(x=pa->disprect.xmin; x<pa->disprect.xmax; x++, ro++, rz++, rp++, fcol+=4, offs++) {
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* per pixel fixed seed */
Bugfix, own collection: Random seeding is still not perfect in render, especially lack of good thread support still. - VectorBlur node was calling seed for each exec, causing other nodes to get fixed random too. - added seed in non-OSA main loop for render - use BLI_srandom, is better than BLI_srand
2006-11-21 15:52:45 +00:00
BLI_thread_srandom(pa->thread, seed++);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
if(*rp) {
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
ps.obi= *ro;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
ps.facenr= *rp;
ps.z= *rz;
if(shade_samples(&ssamp, &ps, x, y)) {
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
/* combined and passes */
add_passes(rl, offs, ssamp.shi, ssamp.shr);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
}
}
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
}
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
if(y&1)
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(R.test_break(R.tbh)) break;
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
}
New shadow feature: Irregular Shadow Buffers Full log: http://www.blender3d.org/cms/Irregular_Shadow_Buffe.785.0.html In short: this is a shadow buffer approach that always results in crispy shadows, independent of lamp buffer size or zoom level. This shadow buffer system also supports transparent shadow. This is part of work on refreshing Shadow Buffers in Blender. You now can choose of two types (Classical, Irregular). More types will follow. Also quality issues for Classical shadow buffers are going to be reviewed, especially to solve the lousy Biasing. For the CVS log record; it is based on articles: Gregory Johnson et al, University of Texas, Austin. (Regular grid method). Timo Aila and Samuli Laine, Helsinki University of Technology. (BSP method).
2006-10-14 10:21:19 +00:00
Approximate Ambient Occlusion ============================= A new approximate ambient occlusion method has been added, next to the existing one based on raytracing. This method is specifically targetted at use in animations, since it is inherently noise free, and so will not flicker across frames. http://www.blender.org/development/current-projects/changes-since-244/approximate-ambient-occlusion/ http://peach.blender.org/index.php/approximate-ambient-occlusion/ Further improvements are still needed, but it can be tested already. There are still a number of known issues: - Bias errors on backfaces. - For performance, instanced object do not occlude currently. - Sky textures don't work well, the derivatives for texture evaluation are not correct. - Multiple passes do not work entirely correct (they are not accurate to begin with, but could be better).
2008-01-17 19:27:16 +00:00
if(R.occlusiontree)
free_occ_samples(&R, pa);
New shadow feature: Irregular Shadow Buffers Full log: http://www.blender3d.org/cms/Irregular_Shadow_Buffe.785.0.html In short: this is a shadow buffer approach that always results in crispy shadows, independent of lamp buffer size or zoom level. This shadow buffer system also supports transparent shadow. This is part of work on refreshing Shadow Buffers in Blender. You now can choose of two types (Classical, Irregular). More types will follow. Also quality issues for Classical shadow buffers are going to be reviewed, especially to solve the lousy Biasing. For the CVS log record; it is based on articles: Gregory Johnson et al, University of Texas, Austin. (Regular grid method). Timo Aila and Samuli Laine, Helsinki University of Technology. (BSP method).
2006-10-14 10:21:19 +00:00
if(R.r.mode & R_SHADOW)
ISB_free(pa);
Initial revision
2002-10-12 11:37:38 +00:00
}
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
/* disable scanline updating */
rr->renlay= NULL;
Initial revision
2002-10-12 11:37:38 +00:00
}
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
Tweaks in RenderLayer rendering: - LampHalos can be rendered separately too. Just disable 'Solid' in a layer and keep 'Halo' option enabled. - Note that disabling 'Solid' will still fill in Z values for the solid faces, to provide occlusion information for the Ztransp and Halo layer options. The latter didn't work this way until now for OSA render. ALso note that that Ztransp+LampHalo still isn't good marriage... it renders a bit weird, but that's an old issue. :)
2006-02-09 22:33:49 +00:00
/* lamphalo after solid, before ztra, looks nicest because ztra does own halo */
if(R.flag & R_LAMPHALO)
if(rl->layflag & SCE_LAY_HALO)
Numerous fixes in Render code: - Bug: material emit was ignored (showed in preview render backdrop) - Bug: world exposure was ignored - Bug: lamp halo was ignoring 'render layer light override'. Further reshuffled the way shadows are being pre-calculated, this to enable more advanced (and faster) usage of Material lightgroups. Now shadows are being cached in lamps, using a per-sample counter to check if a recalc is needed. Will also work (later) for Raytracing node shaders. - New: Material LightGroup option "Always", which always shades the lights in the group, independent of visibility layer. (so it allows to move such lights to hidden layer, not influencing anything).
2006-12-08 09:40:44 +00:00
lamphalo_tile(pa, rl);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
Fix #4087 Flipped layer render order for halo and ztransp. This because the ztransp layer changes z values now (for vblur). This means the default order is: solid + lamphalo + halo + ztra. Different combinations can be achieved with composting nodes.
2006-05-14 09:02:50 +00:00
/* halo before ztra, because ztra fills in zbuffer now */
if(R.flag & R_HALO)
if(rl->layflag & SCE_LAY_HALO)
Fix for bug #8214: halo didn't render correct with FSA.
2008-02-21 20:58:22 +00:00
halo_tile(pa, rl);
Fix #4087 Flipped layer render order for halo and ztransp. This because the ztransp layer changes z values now (for vblur). This means the default order is: solid + lamphalo + halo + ztra. Different combinations can be achieved with composting nodes.
2006-05-14 09:02:50 +00:00
Strands now mix together correctly with ZTransp. They now also store a list of samples per pixel, and then get shaded together with the ztransp samples. This comes with a slight speed hit, but mainly memory might be a concern. However, testing some peach scenes I haven't problems.
2008-01-28 16:54:52 +00:00
if(R.flag & R_ZTRA || R.totstrand) {
if(rl->layflag & (SCE_LAY_ZTRA|SCE_LAY_STRAND)) {
Hooray! Finally found the dreaded "Opengl crash" the poor orange team suffered for the entire movie. :) It only happened when rendering large frames, using a lot of memory and typically when you also use other software in meantime. Reason: the main thread does the drawing updating, while rendering is still continuing. When using Ztransp, there was a free buffer done when possibly a draw could still be in progress. Only crashed when drawing is slow... explaining why it only showed up in more complex cases.
2006-06-30 14:21:25 +00:00
float *fcol, *acol;
Tweaks in RenderLayer rendering: - LampHalos can be rendered separately too. Just disable 'Solid' in a layer and keep 'Halo' option enabled. - Note that disabling 'Solid' will still fill in Z values for the solid faces, to provide occlusion information for the Ztransp and Halo layer options. The latter didn't work this way until now for OSA render. ALso note that that Ztransp+LampHalo still isn't good marriage... it renders a bit weird, but that's an old issue. :)
2006-02-09 22:33:49 +00:00
int x;
Hooray! Finally found the dreaded "Opengl crash" the poor orange team suffered for the entire movie. :) It only happened when rendering large frames, using a lot of memory and typically when you also use other software in meantime. Reason: the main thread does the drawing updating, while rendering is still continuing. When using Ztransp, there was a free buffer done when possibly a draw could still be in progress. Only crashed when drawing is slow... explaining why it only showed up in more complex cases.
2006-06-30 14:21:25 +00:00
/* allocate, but not free here, for asynchronous display of this rect in main thread */
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
rl->acolrect= MEM_callocN(4*sizeof(float)*pa->rectx*pa->recty, "alpha layer");
Hooray! Finally found the dreaded "Opengl crash" the poor orange team suffered for the entire movie. :) It only happened when rendering large frames, using a lot of memory and typically when you also use other software in meantime. Reason: the main thread does the drawing updating, while rendering is still continuing. When using Ztransp, there was a free buffer done when possibly a draw could still be in progress. Only crashed when drawing is slow... explaining why it only showed up in more complex cases.
2006-06-30 14:21:25 +00:00
Tweaks in RenderLayer rendering: - LampHalos can be rendered separately too. Just disable 'Solid' in a layer and keep 'Halo' option enabled. - Note that disabling 'Solid' will still fill in Z values for the solid faces, to provide occlusion information for the Ztransp and Halo layer options. The latter didn't work this way until now for OSA render. ALso note that that Ztransp+LampHalo still isn't good marriage... it renders a bit weird, but that's an old issue. :)
2006-02-09 22:33:49 +00:00
/* swap for live updates */
Hooray! Finally found the dreaded "Opengl crash" the poor orange team suffered for the entire movie. :) It only happened when rendering large frames, using a lot of memory and typically when you also use other software in meantime. Reason: the main thread does the drawing updating, while rendering is still continuing. When using Ztransp, there was a free buffer done when possibly a draw could still be in progress. Only crashed when drawing is slow... explaining why it only showed up in more complex cases.
2006-06-30 14:21:25 +00:00
SWAP(float *, rl->acolrect, rl->rectf);
Bugfix to filter strands with solid correctly.
2008-01-23 20:54:44 +00:00
zbuffer_transp_shade(pa, rl, rl->rectf, NULL);
Hooray! Finally found the dreaded "Opengl crash" the poor orange team suffered for the entire movie. :) It only happened when rendering large frames, using a lot of memory and typically when you also use other software in meantime. Reason: the main thread does the drawing updating, while rendering is still continuing. When using Ztransp, there was a free buffer done when possibly a draw could still be in progress. Only crashed when drawing is slow... explaining why it only showed up in more complex cases.
2006-06-30 14:21:25 +00:00
SWAP(float *, rl->acolrect, rl->rectf);
Tweaks in RenderLayer rendering: - LampHalos can be rendered separately too. Just disable 'Solid' in a layer and keep 'Halo' option enabled. - Note that disabling 'Solid' will still fill in Z values for the solid faces, to provide occlusion information for the Ztransp and Halo layer options. The latter didn't work this way until now for OSA render. ALso note that that Ztransp+LampHalo still isn't good marriage... it renders a bit weird, but that's an old issue. :)
2006-02-09 22:33:49 +00:00
Hooray! Finally found the dreaded "Opengl crash" the poor orange team suffered for the entire movie. :) It only happened when rendering large frames, using a lot of memory and typically when you also use other software in meantime. Reason: the main thread does the drawing updating, while rendering is still continuing. When using Ztransp, there was a free buffer done when possibly a draw could still be in progress. Only crashed when drawing is slow... explaining why it only showed up in more complex cases.
2006-06-30 14:21:25 +00:00
fcol= rl->rectf; acol= rl->acolrect;
Tweaks in RenderLayer rendering: - LampHalos can be rendered separately too. Just disable 'Solid' in a layer and keep 'Halo' option enabled. - Note that disabling 'Solid' will still fill in Z values for the solid faces, to provide occlusion information for the Ztransp and Halo layer options. The latter didn't work this way until now for OSA render. ALso note that that Ztransp+LampHalo still isn't good marriage... it renders a bit weird, but that's an old issue. :)
2006-02-09 22:33:49 +00:00
for(x=pa->rectx*pa->recty; x>0; x--, acol+=4, fcol+=4) {
addAlphaOverFloat(fcol, acol);
Three features; - Live scanline updates while rendering Using a timer system, each second now the tiles that are being processed are checked if they could use display. To make this work pretty, I had to use the threaded 'tile processor' for a single thread too, but that's now proven to be stable. Also note that these updates draw per layer, including ztransp progress separately from solid render. - Recode of ztransp OSA Until now (since blender 1.0) the ztransp part was fully rendered and added on top of the solid part with alpha-over. This adding was done before the solid part applied sub-pixel sample filtering, causing the ztransp layer to be always too blurry. Now the ztransp layer uses same sub=pixel filter, resulting in the same AA level (and filter results) as the solid part. Quite noticable with hair renders. - Vector buffer support & preliminary vector-blur Node Using the "Render Layer" panel "Vector" pass button, the motion vectors per pixel are calculated and stored. Accessible via the Compositor. The vector-blur node is horrible btw! It just uses the length of the vector to apply a filter like with current (z)blur. I'm committing it anyway, I'll experiment with it further, and who knows some surprise code shows up!
2006-01-31 21:49:05 +00:00
}
}
}
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
Bugfix #17656 - Changed order for applying atmosphere, it does it now before alpha-adding sky, giving correct transparency - Added correction for de-premulling and premulling scatter color
2008-09-21 12:03:41 +00:00
/* sun/sky */
if(rl->layflag & SCE_LAY_SKY)
atm_tile(pa, rl);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* sky before edge */
if(rl->layflag & SCE_LAY_SKY)
New rendering option: FSA! This completes the pipeline make-over, as started in 2006. With this option, during rendering, each sample for every layer and pass is being saved on disk (looks like non-antialiased images). Then the composite and color correction happens, then a clip to 0-1 range, and only in end all samples get combined - using sampling filters such as gauss/mitch/catmul. This results in artefact-free antialiased images. Even Z-combine or ID masks now work perfect for it! This is an unfinished commit btw; Brecht will finish this for strands. Also Halo doesnt work yet. To activate FSA: press "Save Buffers" and the new button next to it. :)
2008-01-28 16:33:59 +00:00
sky_tile(pa, rl);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(!R.test_break(R.tbh)) {
Cleanup of UI for new renderpipeline - removed "Unified" button, replaced with "HD" preset for 1920x1080 output - removed the unused "Pass" options - removed the unused "Strands" render-layer option Because the internal render pipe supports this already; added two more render-layer options: - "Sky", to enable/disable sky render in a layer (this was part of "Solid" before, not so correct... to ensure previously saved files work, the "Sky" option is set by default when "Solid" was set. The version patching will do this temporally always, until we've bumped up version to 2.42 - "Edge", to enable/disable edge render in a layer. Nice for compositing. Also in this commit: fixed warnings for exported functions for the new Node Editor pull-down menus.
2006-06-17 10:25:07 +00:00
if(rl->layflag & SCE_LAY_EDGE)
if(R.r.mode & R_EDGE)
edge_enhance_add(pa, rl->rectf, edgerect);
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
}
Orange; more render & compo stuff! -> Rendering in RenderLayers It's important to distinguish a 'render layer' from a 'pass'. The first is control over the main pipeline itself, to indicate what geometry is being is rendered. The 'pass' (not in this commit!) is related to internal shading code, like shadow/spec/AO/normals/etc. Options for RenderLayers now are: - Indicate which 3d 'view layers' have to be included (so you can render front and back separately) - "Solid", all solid faces, includes sky at the moment too - "ZTransp", all transparent faces - "Halo", the halos - "Strand", the particle strands (not coded yet...) Currently only 2 'passes' are exported for render, which is the "Combined" buffer and the "Z. The latter now works, and can be turned on/off. Note that all layers are still fully kept in memory now, saving the tiles and layers to disk (in exr) is also todo. -> New Blur options The existing Blur Node (compositor) now has an optional input image. This has to be a 'value buffer', which can be a Zbuffer, or any mask you can think of. The input values have to be in the 0-1 range, so another new node was added too "Map Value". The value input can also be used to tweak blur size with the (todo) Time Node. Temporal screenies: http://www.blender.org/bf/rt.jpg http://www.blender.org/bf/rt1.jpg http://www.blender.org/bf/rt2.jpg BTW: The compositor is very slow still, it recalulates all nodes on each change still. Persistant memory and dependency checks is coming!
2006-01-26 22:18:46 +00:00
bugfix #5755 Render: Vector blur didn't work proper for Ztransp (accidentally was antialising speed, which should not happen). Also added speed-pass clearing in render, this to elimate the 'tsk tsk' print in the Vector Blur node. :)
2007-01-21 10:04:50 +00:00
if(rl->passflag & SCE_PASS_VECTOR)
reset_sky_speed(pa, rl);
Bugfix, IRC report: Render option "Key Alpha" did not work yet.
2006-06-29 17:22:29 +00:00
/* de-premul alpha */
if(R.r.alphamode & R_ALPHAKEY)
Bugfix #17497 FSA didn't support "key alpha", fixed it. Also noticed edge-enhance didnt work for FSA, added warning for that, fixing it is too clumsy, this option belongs in compositor.
2008-09-07 13:15:44 +00:00
convert_to_key_alpha(pa, rl);
Bugfix, IRC report: Render option "Key Alpha" did not work yet.
2006-06-29 17:22:29 +00:00
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
if(edgerect) MEM_freeN(edgerect);
Bug 4429 When using multiple RenderLayers, each with own 'edge render', the buffer holding edge pixels wasn't cleared from scratch, so the edges of previous layers were showing too.
2006-06-23 16:48:28 +00:00
edgerect= NULL;
Negative zmask rendering: now there is a setting to render everything behind the zmask instead of in front. Might need tweaks, and doesn't work with halo yet.
2008-02-20 14:17:53 +00:00
if(pa->rectmask) {
MEM_freeN(pa->rectmask);
pa->rectmask= NULL;
}
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
}
Three feats; - RenderLayers with 'view layers' set, now also take visible lights into account. Works just like for scene layer settings. - On ESC from render, compositing (if set) is being skipped too - While rendering with multiple RenderLayers it will end with a display of the current RenderLayer (as in Scene buttons)
2006-02-02 21:08:39 +00:00
/* display active layer */
Carefully went over all scanline updating while rendering, to ensure only updates are allowed to draw when a part is within a scanline rendering loop. Might solve threads issues with opengl...
2006-02-21 10:47:08 +00:00
rr->renrect.ymin=rr->renrect.ymax= 0;
Rendering tricks for improving workflow: - Button option "Single" in render-layer panel will enable to only render the currently indicated render-layer. It will also skip compositing. - Brought back the 'Local View' render. This will only render the visible objects, but with lights from the original view-layers. To make the option useful, it also temporal enables 'Single', which has the a disadvantage that you need to set the correct render-layer. It is a bit a tricky option though... since its quite invisble and confusing for people who don't know the feature. This might become either a button in 3d header, or use a popup requester to confirm, or... will need to think over! At least; both options display in render window a text to denote the option.
2006-02-16 12:27:46 +00:00
rr->renlay= render_get_active_layer(&R, rr);
Three feats; - RenderLayers with 'view layers' set, now also take visible lights into account. Works just like for scene layer settings. - On ESC from render, compositing (if set) is being skipped too - While rendering with multiple RenderLayers it will end with a display of the current RenderLayer (as in Scene buttons)
2006-02-02 21:08:39 +00:00
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
MEM_freeN(pa->recto); pa->recto= NULL;
Added support for threadsafe MEM_mallocN/MEM_freeN in the guardedalloc module itself, replacing the special MEM_mallocT/MEM_freeT functions. Mutex locking is only enabled when threads are running. There was no good reason to have these separate, it just led to ugly hacks when calling functions with non-threadsafe malloc from threads.
2006-09-06 19:13:23 +00:00
MEM_freeN(pa->rectp); pa->rectp= NULL;
MEM_freeN(pa->rectz); pa->rectz= NULL;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
/* SSS preprocess tile render, fully threadable */
typedef struct ZBufSSSHandle {
RenderPart *pa;
ListBase psmlist;
int totps;
} ZBufSSSHandle;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
static void addps_sss(void *cb_handle, int obi, int facenr, int x, int y, int z)
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
{
ZBufSSSHandle *handle = cb_handle;
RenderPart *pa= handle->pa;
SSS fixes: - Radius R, G, B sliders had too small number increase on clicking. - Preview render now renders with higher SSS error setting to speed it up a bit. - bug #6664: 3d preview render had artifacts. re->viewdx/dy wasn't set then, which is needed to estimate the area of each point. Have set this now, not in the nicest way, there is some bit duplicated code, but I don't want to refactor existing code with the chance of breaking it at this point. - bug #6665: grid like artifacts with parts rendering. The two extra pixels around parts used for filtering were used as well, leading to double points.
2007-05-08 10:15:51 +00:00
/* extra border for filter gives double samples on part edges,
don't use those */
if(x<pa->crop || x>=pa->rectx-pa->crop)
return;
if(y<pa->crop || y>=pa->recty-pa->crop)
return;
if(pa->rectall) {
Win64: please check my changes if you ran across them ;) But should be fine since no additional crashes were reported!
2008-08-17 17:08:00 +00:00
intptr_t *rs= pa->rectall + pa->rectx*y + x;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
Negative zmask rendering: now there is a setting to render everything behind the zmask instead of in front. Might need tweaks, and doesn't work with halo yet.
2008-02-20 14:17:53 +00:00
addps(&handle->psmlist, rs, obi, facenr, z, 0, 0);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
handle->totps++;
}
SSS fixes: - Radius R, G, B sliders had too small number increase on clicking. - Preview render now renders with higher SSS error setting to speed it up a bit. - bug #6664: 3d preview render had artifacts. re->viewdx/dy wasn't set then, which is needed to estimate the area of each point. Have set this now, not in the nicest way, there is some bit duplicated code, but I don't want to refactor existing code with the chance of breaking it at this point. - bug #6665: grid like artifacts with parts rendering. The two extra pixels around parts used for filtering were used as well, leading to double points.
2007-05-08 10:15:51 +00:00
if(pa->rectz) {
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
int *rz= pa->rectz + pa->rectx*y + x;
int *rp= pa->rectp + pa->rectx*y + x;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
int *ro= pa->recto + pa->rectx*y + x;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
SSS fixes: - Radius R, G, B sliders had too small number increase on clicking. - Preview render now renders with higher SSS error setting to speed it up a bit. - bug #6664: 3d preview render had artifacts. re->viewdx/dy wasn't set then, which is needed to estimate the area of each point. Have set this now, not in the nicest way, there is some bit duplicated code, but I don't want to refactor existing code with the chance of breaking it at this point. - bug #6665: grid like artifacts with parts rendering. The two extra pixels around parts used for filtering were used as well, leading to double points.
2007-05-08 10:15:51 +00:00
if(z < *rz) {
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
if(*rp == 0)
handle->totps++;
*rz= z;
*rp= facenr;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
*ro= obi;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
}
}
SSS fixes: - Radius R, G, B sliders had too small number increase on clicking. - Preview render now renders with higher SSS error setting to speed it up a bit. - bug #6664: 3d preview render had artifacts. re->viewdx/dy wasn't set then, which is needed to estimate the area of each point. Have set this now, not in the nicest way, there is some bit duplicated code, but I don't want to refactor existing code with the chance of breaking it at this point. - bug #6665: grid like artifacts with parts rendering. The two extra pixels around parts used for filtering were used as well, leading to double points.
2007-05-08 10:15:51 +00:00
if(pa->rectbackz) {
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
int *rz= pa->rectbackz + pa->rectx*y + x;
int *rp= pa->rectbackp + pa->rectx*y + x;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
int *ro= pa->rectbacko + pa->rectx*y + x;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
SSS fixes: - Radius R, G, B sliders had too small number increase on clicking. - Preview render now renders with higher SSS error setting to speed it up a bit. - bug #6664: 3d preview render had artifacts. re->viewdx/dy wasn't set then, which is needed to estimate the area of each point. Have set this now, not in the nicest way, there is some bit duplicated code, but I don't want to refactor existing code with the chance of breaking it at this point. - bug #6665: grid like artifacts with parts rendering. The two extra pixels around parts used for filtering were used as well, leading to double points.
2007-05-08 10:15:51 +00:00
if(z >= *rz) {
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
if(*rp == 0)
handle->totps++;
*rz= z;
*rp= facenr;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
*ro= obi;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
}
}
}
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
static void shade_sample_sss(ShadeSample *ssamp, Material *mat, ObjectInstanceRen *obi, VlakRen *vlr, int quad, float x, float y, float z, float *co, float *color, float *area)
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
{
ShadeInput *shi= ssamp->shi;
ShadeResult shr;
Fix for bug #13363: ray (qmc) shadows had some light leaking issues, due to jittering of the start position for antialiasing in a pixel. Now it distributes the start position over the fixed osa sample positions, instead of of random positions in space. The ugly bit is that a custom ordering was defined for osa 8/11/16 to ensure that the first 4 are distributed relatively fair for adaptive sampling to decide if more samples need to be taken.
2008-10-07 15:01:44 +00:00
float texfac, orthoarea, nor[3], alpha, sx, sy;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
/* cache for shadow */
Bugfix: the shadow samplenr increment to reuse shadow results in the renderer could lead to results being reused for unrelated points, result was one wrong pixel or strand per part. Now instead of setting the sample counter to 0 multiple times, it keeps a global counter per thread for the whole render.
2008-03-04 19:58:22 +00:00
shi->samplenr= R.shadowsamplenr[shi->thread]++;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
if(quad)
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
shade_input_set_triangle_i(shi, obi, vlr, 0, 2, 3);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
else
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
shade_input_set_triangle_i(shi, obi, vlr, 0, 1, 2);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
Assorted SSS fixes: - Bug #6652: SSS artifacts with ray shadow. - Bug #6643: AO + SSS crash. couldn't redo this crash myself, but did find a bug that may cause a crash, please test. - Tweak backscattering to avoid thin surfaces rendering too dark.
2007-05-06 05:04:07 +00:00
/* center pixel */
Fix for bug #13363: ray (qmc) shadows had some light leaking issues, due to jittering of the start position for antialiasing in a pixel. Now it distributes the start position over the fixed osa sample positions, instead of of random positions in space. The ugly bit is that a custom ordering was defined for osa 8/11/16 to ensure that the first 4 are distributed relatively fair for adaptive sampling to decide if more samples need to be taken.
2008-10-07 15:01:44 +00:00
sx = x + 0.5f;
sy = y + 0.5f;
Assorted SSS fixes: - Bug #6652: SSS artifacts with ray shadow. - Bug #6643: AO + SSS crash. couldn't redo this crash myself, but did find a bug that may cause a crash, please test. - Tweak backscattering to avoid thin surfaces rendering too dark.
2007-05-06 05:04:07 +00:00
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
/* we estimate the area here using shi->dxco and shi->dyco. we need to
enabled shi->osatex these are filled. we compute two areas, one with
the normal pointed at the camera and one with the original normal, and
then clamp to avoid a too large contribution from a single pixel */
shi->osatex= 1;
VECCOPY(nor, shi->facenor);
Fix for bug #13363: ray (qmc) shadows had some light leaking issues, due to jittering of the start position for antialiasing in a pixel. Now it distributes the start position over the fixed osa sample positions, instead of of random positions in space. The ugly bit is that a custom ordering was defined for osa 8/11/16 to ensure that the first 4 are distributed relatively fair for adaptive sampling to decide if more samples need to be taken.
2008-10-07 15:01:44 +00:00
calc_view_vector(shi->facenor, sx, sy);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
Normalize(shi->facenor);
Fix for bug #13363: ray (qmc) shadows had some light leaking issues, due to jittering of the start position for antialiasing in a pixel. Now it distributes the start position over the fixed osa sample positions, instead of of random positions in space. The ugly bit is that a custom ordering was defined for osa 8/11/16 to ensure that the first 4 are distributed relatively fair for adaptive sampling to decide if more samples need to be taken.
2008-10-07 15:01:44 +00:00
shade_input_set_viewco(shi, x, y, sx, sy, z);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
orthoarea= VecLength(shi->dxco)*VecLength(shi->dyco);
VECCOPY(shi->facenor, nor);
Fix for bug #13363: ray (qmc) shadows had some light leaking issues, due to jittering of the start position for antialiasing in a pixel. Now it distributes the start position over the fixed osa sample positions, instead of of random positions in space. The ugly bit is that a custom ordering was defined for osa 8/11/16 to ensure that the first 4 are distributed relatively fair for adaptive sampling to decide if more samples need to be taken.
2008-10-07 15:01:44 +00:00
shade_input_set_viewco(shi, x, y, sx, sy, z);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
*area= VecLength(shi->dxco)*VecLength(shi->dyco);
*area= MIN2(*area, 2.0f*orthoarea);
shade_input_set_uv(shi);
shade_input_set_normals(shi);
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
/* we don't want flipped normals, they screw up back scattering */
if(shi->flippednor)
shade_input_flip_normals(shi);
Fix for bugs #6709 and #7005: - SSS not working with negative scale objects. - Fluids + SSS not always rendering smooth normals correct.
2007-11-23 14:10:05 +00:00
/* not a pretty solution, but fixes common cases */
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
if(shi->obr->ob && shi->obr->ob->transflag & OB_NEG_SCALE) {
2.50: svn merge https://svn.blender.org/svnroot/bf-blender/trunk/blender -r17853:HEAD
2009-01-26 08:34:40 +00:00
VecNegf(shi->vn);
VecNegf(shi->vno);
Fix for bugs #6709 and #7005: - SSS not working with negative scale objects. - Fluids + SSS not always rendering smooth normals correct.
2007-11-23 14:10:05 +00:00
}
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
/* if nodetree, use the material that we are currently preprocessing
instead of the node material */
if(shi->mat->nodetree && shi->mat->use_nodes)
shi->mat= mat;
/* init material vars */
// note, keep this synced with render_types.h
memcpy(&shi->r, &shi->mat->r, 23*sizeof(float));
shi->har= shi->mat->har;
/* render */
shade_input_set_shade_texco(shi);
shade_samples_do_AO(ssamp);
shade_material_loop(shi, &shr);
VECCOPY(co, shi->co);
VECCOPY(color, shr.combined);
/* texture blending */
texfac= shi->mat->sss_texfac;
Fix for bug #8963: there was still an issue with SSS and transparency in the shading because of ramps.
2008-04-30 12:02:09 +00:00
alpha= shr.combined[3];
Fix or bug #8712: transparency + sss gave too bright results.
2008-04-12 16:31:29 +00:00
*area *= alpha;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
}
Assorted SSS fixes: - Bug #6652: SSS artifacts with ray shadow. - Bug #6643: AO + SSS crash. couldn't redo this crash myself, but did find a bug that may cause a crash, please test. - Tweak backscattering to avoid thin surfaces rendering too dark.
2007-05-06 05:04:07 +00:00
static void zbufshade_sss_free(RenderPart *pa)
{
#if 0
MEM_freeN(pa->rectall); pa->rectall= NULL;
freeps(&handle.psmlist);
#else
MEM_freeN(pa->rectz); pa->rectz= NULL;
MEM_freeN(pa->rectp); pa->rectp= NULL;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
MEM_freeN(pa->recto); pa->recto= NULL;
Assorted SSS fixes: - Bug #6652: SSS artifacts with ray shadow. - Bug #6643: AO + SSS crash. couldn't redo this crash myself, but did find a bug that may cause a crash, please test. - Tweak backscattering to avoid thin surfaces rendering too dark.
2007-05-06 05:04:07 +00:00
MEM_freeN(pa->rectbackz); pa->rectbackz= NULL;
MEM_freeN(pa->rectbackp); pa->rectbackp= NULL;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
MEM_freeN(pa->rectbacko); pa->rectbacko= NULL;
Assorted SSS fixes: - Bug #6652: SSS artifacts with ray shadow. - Bug #6643: AO + SSS crash. couldn't redo this crash myself, but did find a bug that may cause a crash, please test. - Tweak backscattering to avoid thin surfaces rendering too dark.
2007-05-06 05:04:07 +00:00
#endif
}
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
void zbufshade_sss_tile(RenderPart *pa)
{
Fix preview rendering for SSS to not show preprocessing pass that leads to flickering, and a bug that would display parts of the preview render as black for interrupted preview renders. (sorry for the late fix, i've triple checked the code ..)
2007-05-09 16:25:58 +00:00
Render *re= &R;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
ShadeSample ssamp;
ZBufSSSHandle handle;
RenderResult *rr= pa->result;
Fixes for SSS with render layers. Now in the preprocessing pass it uses flags from all render layers added together, not fully correct yet, but it's not so easy to separate passes cleanly with SSS.
2008-01-24 15:41:21 +00:00
RenderLayer *rl;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
VlakRen *vlr;
Fix preview rendering for SSS to not show preprocessing pass that leads to flickering, and a bug that would display parts of the preview render as black for interrupted preview renders. (sorry for the late fix, i've triple checked the code ..)
2007-05-09 16:25:58 +00:00
Material *mat= re->sss_mat;
Fixes for SSS with render layers. Now in the preprocessing pass it uses flags from all render layers added together, not fully correct yet, but it's not so easy to separate passes cleanly with SSS.
2008-01-24 15:41:21 +00:00
float (*co)[3], (*color)[3], *area, *fcol;
Fix preview rendering for SSS to not show preprocessing pass that leads to flickering, and a bug that would display parts of the preview render as black for interrupted preview renders. (sorry for the late fix, i've triple checked the code ..)
2007-05-09 16:25:58 +00:00
int x, y, seed, quad, totpoint, display = !(re->r.scemode & R_PREVIEWBUTS);
Fixes for SSS with render layers. Now in the preprocessing pass it uses flags from all render layers added together, not fully correct yet, but it's not so easy to separate passes cleanly with SSS.
2008-01-24 15:41:21 +00:00
int *ro, *rz, *rp, *rbo, *rbz, *rbp, lay;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
#if 0
PixStr *ps;
Win64: please check my changes if you ran across them ;) But should be fine since no additional crashes were reported!
2008-08-17 17:08:00 +00:00
intptr_t *rs;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
int z;
#endif
/* setup pixelstr list and buffer for zbuffering */
handle.pa= pa;
handle.totps= 0;
#if 0
handle.psmlist.first= handle.psmlist.last= NULL;
addpsmain(&handle.psmlist);
Win64: please check my changes if you ran across them ;) But should be fine since no additional crashes were reported!
2008-08-17 17:08:00 +00:00
pa->rectall= MEM_callocN(sizeof(intptr_t)*pa->rectx*pa->recty+4, "rectall");
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
#else
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
pa->recto= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "recto");
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
pa->rectp= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "rectp");
pa->rectz= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "rectz");
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
pa->rectbacko= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "rectbacko");
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
pa->rectbackp= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "rectbackp");
pa->rectbackz= MEM_mallocN(sizeof(int)*pa->rectx*pa->recty, "rectbackz");
#endif
Fixes for SSS with render layers. Now in the preprocessing pass it uses flags from all render layers added together, not fully correct yet, but it's not so easy to separate passes cleanly with SSS.
2008-01-24 15:41:21 +00:00
/* setup shade sample with correct passes */
memset(&ssamp, 0, sizeof(ssamp));
shade_sample_initialize(&ssamp, pa, rr->layers.first);
ssamp.tot= 1;
for(rl=rr->layers.first; rl; rl=rl->next) {
ssamp.shi[0].lay |= rl->lay;
ssamp.shi[0].layflag |= rl->layflag;
ssamp.shi[0].passflag |= rl->passflag;
ssamp.shi[0].combinedflag |= ~rl->pass_xor;
}
rl= rr->layers.first;
ssamp.shi[0].passflag |= SCE_PASS_RGBA|SCE_PASS_COMBINED;
ssamp.shi[0].combinedflag &= ~(SCE_PASS_SPEC);
Fix for bug #9081: SSS preprocess should not use material or light overrides, since it works independent of render layers.
2008-04-19 21:59:15 +00:00
ssamp.shi[0].mat_override= NULL;
ssamp.shi[0].light_override= NULL;
Fixes for SSS with render layers. Now in the preprocessing pass it uses flags from all render layers added together, not fully correct yet, but it's not so easy to separate passes cleanly with SSS.
2008-01-24 15:41:21 +00:00
lay= ssamp.shi[0].lay;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
/* create the pixelstrs to be used later */
Fixes for SSS with render layers. Now in the preprocessing pass it uses flags from all render layers added together, not fully correct yet, but it's not so easy to separate passes cleanly with SSS.
2008-01-24 15:41:21 +00:00
zbuffer_sss(pa, lay, &handle, addps_sss);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
Assorted SSS fixes: - Bug #6652: SSS artifacts with ray shadow. - Bug #6643: AO + SSS crash. couldn't redo this crash myself, but did find a bug that may cause a crash, please test. - Tweak backscattering to avoid thin surfaces rendering too dark.
2007-05-06 05:04:07 +00:00
if(handle.totps==0) {
zbufshade_sss_free(pa);
return;
}
Fixes for SSS with render layers. Now in the preprocessing pass it uses flags from all render layers added together, not fully correct yet, but it's not so easy to separate passes cleanly with SSS.
2008-01-24 15:41:21 +00:00
fcol= rl->rectf;
Assorted SSS fixes: - Bug #6652: SSS artifacts with ray shadow. - Bug #6643: AO + SSS crash. couldn't redo this crash myself, but did find a bug that may cause a crash, please test. - Tweak backscattering to avoid thin surfaces rendering too dark.
2007-05-06 05:04:07 +00:00
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
co= MEM_mallocN(sizeof(float)*3*handle.totps, "SSSCo");
color= MEM_mallocN(sizeof(float)*3*handle.totps, "SSSColor");
area= MEM_mallocN(sizeof(float)*handle.totps, "SSSArea");
#if 0
/* create ISB (does not work currently!) */
Fix preview rendering for SSS to not show preprocessing pass that leads to flickering, and a bug that would display parts of the preview render as black for interrupted preview renders. (sorry for the late fix, i've triple checked the code ..)
2007-05-09 16:25:58 +00:00
if(re->r.mode & R_SHADOW)
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
ISB_create(pa, NULL);
#endif
Fix preview rendering for SSS to not show preprocessing pass that leads to flickering, and a bug that would display parts of the preview render as black for interrupted preview renders. (sorry for the late fix, i've triple checked the code ..)
2007-05-09 16:25:58 +00:00
if(display) {
/* initialize scanline updates for main thread */
rr->renrect.ymin= 0;
rr->renlay= rl;
}
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
seed= pa->rectx*pa->disprect.ymin;
#if 0
rs= pa->rectall;
#else
rz= pa->rectz;
rp= pa->rectp;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
ro= pa->recto;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
rbz= pa->rectbackz;
rbp= pa->rectbackp;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
rbo= pa->rectbacko;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
#endif
totpoint= 0;
for(y=pa->disprect.ymin; y<pa->disprect.ymax; y++, rr->renrect.ymax++) {
for(x=pa->disprect.xmin; x<pa->disprect.xmax; x++, fcol+=4) {
/* per pixel fixed seed */
BLI_thread_srandom(pa->thread, seed++);
#if 0
if(rs) {
/* for each sample in this pixel, shade it */
for(ps=(PixStr*)*rs; ps; ps=ps->next) {
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
ObjectInstanceRen *obi= &re->objectinstance[ps->obi];
ObjectRen *obr= obi->obr;
vlr= RE_findOrAddVlak(obr, (ps->facenr-1) & RE_QUAD_MASK);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
quad= (ps->facenr & RE_QUAD_OFFS);
z= ps->z;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
shade_sample_sss(&ssamp, mat, obi, vlr, quad, x, y, z,
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
co[totpoint], color[totpoint], &area[totpoint]);
totpoint++;
VECADD(fcol, fcol, color);
fcol[3]= 1.0f;
}
rs++;
}
#else
if(rp) {
if(*rp != 0) {
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
ObjectInstanceRen *obi= &re->objectinstance[*ro];
ObjectRen *obr= obi->obr;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
/* shade front */
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
vlr= RE_findOrAddVlak(obr, (*rp-1) & RE_QUAD_MASK);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
quad= ((*rp) & RE_QUAD_OFFS);
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
shade_sample_sss(&ssamp, mat, obi, vlr, quad, x, y, *rz,
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
co[totpoint], color[totpoint], &area[totpoint]);
Assorted SSS fixes: - Bug #6652: SSS artifacts with ray shadow. - Bug #6643: AO + SSS crash. couldn't redo this crash myself, but did find a bug that may cause a crash, please test. - Tweak backscattering to avoid thin surfaces rendering too dark.
2007-05-06 05:04:07 +00:00
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
VECADD(fcol, fcol, color[totpoint]);
fcol[3]= 1.0f;
totpoint++;
}
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
rp++; rz++; ro++;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
}
if(rbp) {
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
if(*rbp != 0 && !(*rbp == *(rp-1) && *rbo == *(ro-1))) {
ObjectInstanceRen *obi= &re->objectinstance[*rbo];
ObjectRen *obr= obi->obr;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
/* shade back */
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
vlr= RE_findOrAddVlak(obr, (*rbp-1) & RE_QUAD_MASK);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
quad= ((*rbp) & RE_QUAD_OFFS);
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
shade_sample_sss(&ssamp, mat, obi, vlr, quad, x, y, *rbz,
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
co[totpoint], color[totpoint], &area[totpoint]);
/* to indicate this is a back sample */
area[totpoint]= -area[totpoint];
VECADD(fcol, fcol, color[totpoint]);
fcol[3]= 1.0f;
totpoint++;
}
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
rbz++; rbp++; rbo++;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
}
#endif
}
if(y&1)
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(re->test_break(re->tbh)) break;
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
}
/* note: after adding we do not free these arrays, sss keeps them */
if(totpoint > 0) {
Fix preview rendering for SSS to not show preprocessing pass that leads to flickering, and a bug that would display parts of the preview render as black for interrupted preview renders. (sorry for the late fix, i've triple checked the code ..)
2007-05-09 16:25:58 +00:00
sss_add_points(re, co, color, area, totpoint);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
}
else {
MEM_freeN(co);
MEM_freeN(color);
MEM_freeN(area);
}
#if 0
Fix preview rendering for SSS to not show preprocessing pass that leads to flickering, and a bug that would display parts of the preview render as black for interrupted preview renders. (sorry for the late fix, i've triple checked the code ..)
2007-05-09 16:25:58 +00:00
if(re->r.mode & R_SHADOW)
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
ISB_free(pa);
#endif
Fix preview rendering for SSS to not show preprocessing pass that leads to flickering, and a bug that would display parts of the preview render as black for interrupted preview renders. (sorry for the late fix, i've triple checked the code ..)
2007-05-09 16:25:58 +00:00
if(display) {
/* display active layer */
rr->renrect.ymin=rr->renrect.ymax= 0;
rr->renlay= render_get_active_layer(&R, rr);
}
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
Assorted SSS fixes: - Bug #6652: SSS artifacts with ray shadow. - Bug #6643: AO + SSS crash. couldn't redo this crash myself, but did find a bug that may cause a crash, please test. - Tweak backscattering to avoid thin surfaces rendering too dark.
2007-05-06 05:04:07 +00:00
zbufshade_sss_free(pa);
Subsurface scattering: Documentation on the settings, known limitations and implementation info can be found here: http://www.blender.org/development/current-projects/changes-since-243/subsurface-scattering/
2007-05-03 21:37:52 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
/* ------------------------------------------------------------------------ */
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
static void renderhalo_post(RenderResult *rr, float *rectf, HaloRen *har) /* postprocess version */
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
{
float dist, xsq, ysq, xn, yn, colf[4], *rectft, *rtf;
Bugfix #4927 Render: Flares were not corrected yet for border render.
2006-09-10 10:35:28 +00:00
float haloxs, haloys;
Initial revision
2002-10-12 11:37:38 +00:00
int minx, maxx, miny, maxy, x, y;
Bugfix #4927 Render: Flares were not corrected yet for border render.
2006-09-10 10:35:28 +00:00
/* calculate the disprect mapped coordinate for halo. note: rectx is disprect corrected */
haloxs= har->xs - R.disprect.xmin;
haloys= har->ys - R.disprect.ymin;
har->miny= miny= haloys - har->rad/R.ycor;
har->maxy= maxy= haloys + har->rad/R.ycor;
Initial revision
2002-10-12 11:37:38 +00:00
if(maxy<0);
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
else if(rr->recty<miny);
Initial revision
2002-10-12 11:37:38 +00:00
else {
Bugfix #4927 Render: Flares were not corrected yet for border render.
2006-09-10 10:35:28 +00:00
minx= floor(haloxs-har->rad);
maxx= ceil(haloxs+har->rad);
Initial revision
2002-10-12 11:37:38 +00:00
if(maxx<0);
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
else if(rr->rectx<minx);
Initial revision
2002-10-12 11:37:38 +00:00
else {
if(minx<0) minx= 0;
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
if(maxx>=rr->rectx) maxx= rr->rectx-1;
Initial revision
2002-10-12 11:37:38 +00:00
if(miny<0) miny= 0;
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
if(maxy>rr->recty) maxy= rr->recty;
Initial revision
2002-10-12 11:37:38 +00:00
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
rectft= rectf+ 4*rr->rectx*miny;
Render lens flare works again. Didn't work due to wrong value for zbuffer passed on (zbuffer values changed with previous commit) Also fixed previewrender for lens flare.
2005-01-18 15:22:42 +00:00
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
for(y=miny; y<maxy; y++) {
Initial revision
2002-10-12 11:37:38 +00:00
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
rtf= rectft+4*minx;
Bugfix #4927 Render: Flares were not corrected yet for border render.
2006-09-10 10:35:28 +00:00
yn= (y - haloys)*R.ycor;
Initial revision
2002-10-12 11:37:38 +00:00
ysq= yn*yn;
for(x=minx; x<=maxx; x++) {
Bugfix #4927 Render: Flares were not corrected yet for border render.
2006-09-10 10:35:28 +00:00
xn= x - haloxs;
Initial revision
2002-10-12 11:37:38 +00:00
xsq= xn*xn;
dist= xsq+ysq;
if(dist<har->radsq) {
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
Patch #8034: "soft" option for halos, which avoids ugly intersections with geometry, and makes halos look more volumetric. Patch contributed by Markus Ilmola, thanks!
2008-04-14 19:48:14 +00:00
if(shadeHaloFloat(har, colf, 0x7FFFFF, dist, xn, yn, har->flarec))
addalphaAddfacFloat(rtf, colf, har->add);
Initial revision
2002-10-12 11:37:38 +00:00
}
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
rtf+=4;
Initial revision
2002-10-12 11:37:38 +00:00
}
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
rectft+= 4*rr->rectx;
Initial revision
2002-10-12 11:37:38 +00:00
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(R.test_break(R.tbh)) break;
Initial revision
2002-10-12 11:37:38 +00:00
}
}
}
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
/* ------------------------------------------------------------------------ */
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
static void renderflare(RenderResult *rr, float *rectf, HaloRen *har)
Initial revision
2002-10-12 11:37:38 +00:00
{
Phew, a lot of work, and no new features... Main target was to make the inner rendering loop using no globals anymore. This is essential for proper usage while raytracing, it caused a lot of hacks in the raycode as well, which even didn't work correctly for all situations (textures especially). Done this by creating a new local struct RenderInput, which replaces usage of the global struct Render R. The latter now only is used to denote image size, viewmatrix, and the like. Making the inner render loops using no globals caused 1000s of vars to be changed... but the result definitely is much nicer code, which enables making 'real' shaders in a next stage. It also enabled me to remove the hacks from ray.c Then i went to the task of removing redundant code. Especially the calculus of texture coords took place (identical) in three locations. Most obvious is the change in the unified render part, which is much less code now; it uses the same rendering routines as normal render now. (Note; not for halos yet!) I also removed 6 files called 'shadowbuffer' something. This was experimen- tal stuff from NaN days. And again saved a lot of double used code. Finally I went over the blenkernel and blender/src calls to render stuff. Here the same local data is used now, resulting in less dependency. I also moved render-texture to the render module, this was still in Kernel. (new file: texture.c) So! After this commit I will check on the autofiles, to try to fix that. MSVC people have to do it themselves. This commit will need quite some testing help, but I'm around!
2003-12-21 21:52:51 +00:00
extern float hashvectf[];
Initial revision
2002-10-12 11:37:38 +00:00
HaloRen fla;
Material *ma;
float *rc, rad, alfa, visifac, vec[3];
int b, type;
fla= *har;
fla.linec= fla.ringc= fla.flarec= 0;
rad= har->rad;
alfa= har->alfa;
visifac= R.ycor*(har->pixels);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* all radials added / r^3 == 1.0f! */
Initial revision
2002-10-12 11:37:38 +00:00
visifac /= (har->rad*har->rad*har->rad);
visifac*= visifac;
ma= har->mat;
first commit with hours and hours of comment translation for render code...
2003-04-27 14:05:41 +00:00
/* first halo: just do */
Initial revision
2002-10-12 11:37:38 +00:00
har->rad= rad*ma->flaresize*visifac;
har->radsq= har->rad*har->rad;
Render lens flare works again. Didn't work due to wrong value for zbuffer passed on (zbuffer values changed with previous commit) Also fixed previewrender for lens flare.
2005-01-18 15:22:42 +00:00
har->zs= fla.zs= 0;
Initial revision
2002-10-12 11:37:38 +00:00
har->alfa= alfa*visifac;
Render lens flare works again. Didn't work due to wrong value for zbuffer passed on (zbuffer values changed with previous commit) Also fixed previewrender for lens flare.
2005-01-18 15:22:42 +00:00
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
renderhalo_post(rr, rectf, har);
Initial revision
2002-10-12 11:37:38 +00:00
first commit with hours and hours of comment translation for render code...
2003-04-27 14:05:41 +00:00
/* next halo's: the flares */
Initial revision
2002-10-12 11:37:38 +00:00
rc= hashvectf + ma->seed2;
for(b=1; b<har->flarec; b++) {
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
fla.r= fabs(rc[0]);
fla.g= fabs(rc[1]);
fla.b= fabs(rc[2]);
Initial revision
2002-10-12 11:37:38 +00:00
fla.alfa= ma->flareboost*fabs(alfa*visifac*rc[3]);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
fla.hard= 20.0f + fabs(70*rc[7]);
Initial revision
2002-10-12 11:37:38 +00:00
fla.tex= 0;
type= (int)(fabs(3.9*rc[6]));
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
fla.rad= ma->subsize*sqrt(fabs(2.0f*har->rad*rc[4]));
Initial revision
2002-10-12 11:37:38 +00:00
if(type==3) {
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
fla.rad*= 3.0f;
Initial revision
2002-10-12 11:37:38 +00:00
fla.rad+= R.rectx/10;
}
fla.radsq= fla.rad*fla.rad;
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
vec[0]= 1.4*rc[5]*(har->xs-R.winx/2);
vec[1]= 1.4*rc[5]*(har->ys-R.winy/2);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
vec[2]= 32.0f*sqrt(vec[0]*vec[0] + vec[1]*vec[1] + 1.0f);
Initial revision
2002-10-12 11:37:38 +00:00
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
fla.xs= R.winx/2 + vec[0] + (1.2+rc[8])*R.rectx*vec[0]/vec[2];
fla.ys= R.winy/2 + vec[1] + (1.2+rc[8])*R.rectx*vec[1]/vec[2];
Initial revision
2002-10-12 11:37:38 +00:00
if(R.flag & R_SEC_FIELD) {
if(R.r.mode & R_ODDFIELD) fla.ys += 0.5;
else fla.ys -= 0.5;
}
if(type & 1) fla.type= HA_FLARECIRC;
else fla.type= 0;
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
renderhalo_post(rr, rectf, &fla);
Initial revision
2002-10-12 11:37:38 +00:00
fla.alfa*= 0.5;
if(type & 2) fla.type= HA_FLARECIRC;
else fla.type= 0;
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
renderhalo_post(rr, rectf, &fla);
Initial revision
2002-10-12 11:37:38 +00:00
rc+= 7;
}
Giant commit! A full detailed description of this will be done later... is several days of work. Here's a summary: Render: - Full cleanup of render code, removing *all* globals and bad level calls all over blender. Render module is now not called abusive anymore - API-fied calls to rendering - Full recode of internal render pipeline. Is now rendering tiles by default, prepared for much smarter 'bucket' render later. - Each thread now can render a full part - Renders were tested with 4 threads, goes fine, apart from some lookup tables in softshadow and AO still - Rendering is prepared to do multiple layers and passes - No single 32 bits trick in render code anymore, all 100% floats now. Writing images/movies - moved writing images to blender kernel (bye bye 'schrijfplaatje'!) - made a new Movie handle system, also in kernel. This will enable much easier use of movies in Blender PreviewRender: - Using new render API, previewrender (in buttons) now uses regular render code to generate images. - new datafile 'preview.blend.c' has the preview scenes in it - previews get rendered in exact displayed size (1 pixel = 1 pixel) 3D Preview render - new; press Pkey in 3d window, for a panel that continuously renders (pkey is for games, i know... but we dont do that in orange now!) - this render works nearly identical to buttons-preview render, so it stops rendering on any event (mouse, keyboard, etc) - on moving/scaling the panel, the render code doesn't recreate all geometry - same for shifting/panning view - all other operations (now) regenerate the full render database still. - this is WIP... but big fun, especially for simple scenes! Compositor - Using same node system as now in use for shaders, you can composit images - works pretty straightforward... needs much more options/tools and integration with rendering still - is not threaded yet, nor is so smart to only recalculate changes... will be done soon! - the "Render Result" node will get all layers/passes as output sockets - The "Output" node renders to a builtin image, which you can view in the Image window. (yes, output nodes to render-result, and to files, is on the list!) The Bad News - "Unified Render" is removed. It might come back in some stage, but this system should be built from scratch. I can't really understand this code... I expect it is not much needed, especially with advanced layer/passes control - Panorama render, Field render, Motion blur, is not coded yet... (I had to recode every single feature in render, so...!) - Lens Flare is also not back... needs total revision, might become composit effect though (using zbuffer for visibility) - Part render is gone! (well, thats obvious, its default now). - The render window is only restored with limited functionality... I am going to check first the option to render to a Image window, so Blender can become a true single-window application. :) For example, the 'Spare render buffer' (jkey) doesnt work. - Render with border, now default creates a smaller image - No zbuffers are written yet... on the todo! - Scons files and MSVC will need work to get compiling again OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
/* needs recode... integrate this better! */
void add_halo_flare(Render *re)
Initial revision
2002-10-12 11:37:38 +00:00
{
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
RenderResult *rr= re->result;
RenderLayer *rl;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
HaloRen *har;
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
int a, mode, do_draw=0;
/* for now, we get the first renderlayer in list with halos set */
for(rl= rr->layers.first; rl; rl= rl->next)
if(rl->layflag & SCE_LAY_HALO)
break;
if(rl==NULL || rl->rectf==NULL)
return;
Initial revision
2002-10-12 11:37:38 +00:00
mode= R.r.mode;
R.r.mode &= ~R_PANORAMA;
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
Particles ========= - Fix crash in particle transform with the particle system not editable. - Particle child distribution and caching is now multithreaded. - Child particles now have a separate Render Amount next to the existing Amount. The render amount particles are now only distributed and cached at render time, which should make editing with child particles faster. - Two new options for diffuse strand shading: - Surface Diffuse: computes the strand normal taking the normal at the surface into account. - Blending Distance: the distance in Blender units over which to blend in the normal at the surface. - Special strand rendering for more memory efficient and faster hair and grass. This is a work in progress, and has a number of known issues, don't report bugs to me for this feature yet. More info: http://www.blender.org/development/current-projects/changes-since-244/particles/
2007-12-04 13:57:28 +00:00
project_renderdata(&R, projectverto, 0, 0, 0);
Initial revision
2002-10-12 11:37:38 +00:00
for(a=0; a<R.tothalo; a++) {
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
har= R.sortedhalos[a];
Initial revision
2002-10-12 11:37:38 +00:00
if(har->flarec) {
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
do_draw= 1;
renderflare(rr, rl->rectf, har);
Initial revision
2002-10-12 11:37:38 +00:00
}
}
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
if(do_draw) {
/* weak... the display callback wants an active renderlayer pointer... */
rr->renlay= rl;
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
re->display_draw(re->ddh, rr, NULL);
Lens flare rendering back.... completely forgot about this antique feature. It doesn't render preview yet... for that we have to fix preview system.
2006-06-13 14:51:17 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
R.r.mode= mode;
Biiig commit! Thanks to 2-3 weeks of cvs freeze... Render: - New; support for dual CPU render (SDL thread) Currently only works with alternating scanlines, but gives excellent performance. For both normal render as unified implemented. Note the "mutex" locks on z-transp buffer render and imbuf loads. - This has been made possible by major cleanups in render code, especially getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct OSA or using Materials or Texture data to write to. - Made normal render fully 4x32 floats too, and removed all old optimizes with chars or shorts. - Made normal render and unified render use same code for sky and halo render, giving equal (and better) results for halo render. Old render now also uses PostProcess options (brightness, mul, gamma) - Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer after render. Using PostProcess menu you will note an immediate re- display of image too (32 bits RGBA) - Added "Hue" and "Saturation" sliders to PostProcess options - Render module is still not having a "nice" API, but amount of dependencies went down a lot. Next todo: remove abusive "previewrender" code. The last main global in Render (struct Render) now can be re-used for fully controlling a render, to allow multiple "instances" of render to open. - Renderwindow now displays a smal bar on top with the stats, and keeps the stats after render too. Including "spare" page support. Not only easier visible that way, but also to remove the awkward code that was drawing stats in the Info header (extreme slow on some ATIs too) - Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping defines. - I might have forgotten stuff... and will write a nice doc on the architecture!
2004-12-27 19:28:52 +00:00
}
Initial revision
2002-10-12 11:37:38 +00:00
Two fixes: - Shaded view was black, needed to tell it to use 'combined' pass :) - Crash when deleting in compositor all nodes with an output.
2006-12-05 18:02:43 +00:00
/* ************************* used for shaded view ************************ */
More render pipeline finishing up; - Shaded drawmode is back (shift+z). Note it still only uses orco texture; but lighting/shading is using the internal render module entirely. - "Make Sticky" option back. (Also fix in sticky texture render, was wrong scaled)
2006-05-28 11:49:22 +00:00
/* if *re, then initialize, otherwise execute */
void RE_shade_external(Render *re, ShadeInput *shi, ShadeResult *shr)
{
static VlakRen vlr;
Minor render memory usage optimization, removed layer and radface from VlakRen, saves about 100mb for 10 million faces.
2008-01-23 13:35:51 +00:00
static ObjectRen obr;
static ObjectInstanceRen obi;
More render pipeline finishing up; - Shaded drawmode is back (shift+z). Note it still only uses orco texture; but lighting/shading is using the internal render module entirely. - "Make Sticky" option back. (Also fix in sticky texture render, was wrong scaled)
2006-05-28 11:49:22 +00:00
/* init */
if(re) {
R= *re;
/* fake render face */
memset(&vlr, 0, sizeof(VlakRen));
Minor render memory usage optimization, removed layer and radface from VlakRen, saves about 100mb for 10 million faces.
2008-01-23 13:35:51 +00:00
memset(&obr, 0, sizeof(ObjectRen));
memset(&obi, 0, sizeof(ObjectInstanceRen));
obr.lay= -1;
obi.obr= &obr;
More render pipeline finishing up; - Shaded drawmode is back (shift+z). Note it still only uses orco texture; but lighting/shading is using the internal render module entirely. - "Make Sticky" option back. (Also fix in sticky texture render, was wrong scaled)
2006-05-28 11:49:22 +00:00
return;
}
shi->vlr= &vlr;
Minor render memory usage optimization, removed layer and radface from VlakRen, saves about 100mb for 10 million faces.
2008-01-23 13:35:51 +00:00
shi->obr= &obr;
shi->obi= &obi;
More render pipeline finishing up; - Shaded drawmode is back (shift+z). Note it still only uses orco texture; but lighting/shading is using the internal render module entirely. - "Make Sticky" option back. (Also fix in sticky texture render, was wrong scaled)
2006-05-28 11:49:22 +00:00
if(shi->mat->nodetree && shi->mat->use_nodes)
ntreeShaderExecTree(shi->mat->nodetree, shi, shr);
else {
/* copy all relevant material vars, note, keep this synced with render_types.h */
memcpy(&shi->r, &shi->mat->r, 23*sizeof(float));
shi->har= shi->mat->har;
shade_material_loop(shi, shr);
}
}
Two fixes: - Shaded view was black, needed to tell it to use 'combined' pass :) - Crash when deleting in compositor all nodes with an output.
2006-12-05 18:02:43 +00:00
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
/* ************************* bake ************************ */
typedef struct BakeShade {
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
ShadeSample ssamp;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
ObjectInstanceRen *obi;
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
VlakRen *vlr;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
ZSpan *zspan;
Image *ima;
The Big Image refactor! Please read: http://www.blender3d.org/cms/Imaging.834.0.html Or in short: - adding MultiLayer Image support - recoded entire Image API - better integration of movie/sequence Images Was a whole load of work... went down for a week to do this. So, will need a lot of testing! Will be in irc all evening.
2006-12-20 17:57:56 +00:00
ImBuf *ibuf;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
int rectx, recty, quad, type, vdone, ready;
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
float dir[3];
Object *actob;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
unsigned int *rect;
float *rect_float;
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
int usemask;
char *rect_mask; /* bake pixel mask */
Baking now does material texture anti-aliasing if the OSA button is enabled, result of that is mainly visible for image textures.
2008-03-12 20:22:05 +00:00
float dxco[3], dyco[3];
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
} BakeShade;
Initial revision
2002-10-12 11:37:38 +00:00
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
/* bake uses a char mask to know what has been baked */
#define BAKE_MASK_NULL 0
#define BAKE_MASK_MARGIN 1
#define BAKE_MASK_BAKED 2
static void bake_mask_filter_extend( char *mask, int width, int height )
{
char *row1, *row2, *row3;
int rowlen, x, y;
char *temprect;
rowlen= width;
/* make a copy, to prevent flooding */
temprect= MEM_dupallocN(mask);
for(y=1; y<=height; y++) {
/* setup rows */
row1= (char *)(temprect + (y-2)*rowlen);
row2= row1 + rowlen;
row3= row2 + rowlen;
if(y==1)
row1= row2;
else if(y==height)
row3= row2;
for(x=0; x<rowlen; x++) {
if (mask[((y-1)*rowlen)+x]==0) {
if (*row1 || *row2 || *row3 || *(row1+1) || *(row3+1) ) {
mask[((y-1)*rowlen)+x] = BAKE_MASK_MARGIN;
} else if((x!=rowlen-1) && (*(row1+2) || *(row2+2) || *(row3+2)) ) {
mask[((y-1)*rowlen)+x] = BAKE_MASK_MARGIN;
}
}
if(x!=0) {
row1++; row2++; row3++;
}
}
}
MEM_freeN(temprect);
}
static void bake_mask_clear( ImBuf *ibuf, char *mask, char val )
{
int x,y;
if (ibuf->rect_float) {
for(x=0; x<ibuf->x; x++) {
for(y=0; y<ibuf->y; y++) {
if (mask[ibuf->x*y + x] == val) {
float *col= ibuf->rect_float + 4*(ibuf->x*y + x);
col[0] = col[1] = col[2] = col[3] = 0.0f;
}
}
}
} else {
/* char buffer */
for(x=0; x<ibuf->x; x++) {
for(y=0; y<ibuf->y; y++) {
if (mask[ibuf->x*y + x] == val) {
char *col= (char *)(ibuf->rect + ibuf->x*y + x);
col[0] = col[1] = col[2] = col[3] = 0;
}
}
}
}
}
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
static void bake_set_shade_input(ObjectInstanceRen *obi, VlakRen *vlr, ShadeInput *shi, int quad, int isect, int x, int y, float u, float v)
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
{
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
if(isect) {
/* raytrace intersection with different u,v than scanconvert */
if(vlr->v4) {
if(quad)
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
shade_input_set_triangle_i(shi, obi, vlr, 2, 1, 3);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
else
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
shade_input_set_triangle_i(shi, obi, vlr, 0, 1, 3);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
}
else
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
shade_input_set_triangle_i(shi, obi, vlr, 0, 1, 2);
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
}
else {
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
/* regular scanconvert */
if(quad)
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
shade_input_set_triangle_i(shi, obi, vlr, 0, 2, 3);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
else
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
shade_input_set_triangle_i(shi, obi, vlr, 0, 1, 2);
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
}
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
Numerous fixes in Render code: - Bug: material emit was ignored (showed in preview render backdrop) - Bug: world exposure was ignored - Bug: lamp halo was ignoring 'render layer light override'. Further reshuffled the way shadows are being pre-calculated, this to enable more advanced (and faster) usage of Material lightgroups. Now shadows are being cached in lamps, using a per-sample counter to check if a recalc is needed. Will also work (later) for Raytracing node shaders. - New: Material LightGroup option "Always", which always shades the lights in the group, independent of visibility layer. (so it allows to move such lights to hidden layer, not influencing anything).
2006-12-08 09:40:44 +00:00
/* cache for shadow */
Bugfix: the shadow samplenr increment to reuse shadow results in the renderer could lead to results being reused for unrelated points, result was one wrong pixel or strand per part. Now instead of setting the sample counter to 0 multiple times, it keeps a global counter per thread for the whole render.
2008-03-04 19:58:22 +00:00
shi->samplenr= R.shadowsamplenr[shi->thread]++;
Fix for bake texture antialiasing commit, AO with constant jitter didn't work. Also fixed a warning.
2008-03-13 14:36:17 +00:00
shi->mask= 0xFFFF; /* all samples */
Numerous fixes in Render code: - Bug: material emit was ignored (showed in preview render backdrop) - Bug: world exposure was ignored - Bug: lamp halo was ignoring 'render layer light override'. Further reshuffled the way shadows are being pre-calculated, this to enable more advanced (and faster) usage of Material lightgroups. Now shadows are being cached in lamps, using a per-sample counter to check if a recalc is needed. Will also work (later) for Raytracing node shaders. - New: Material LightGroup option "Always", which always shades the lights in the group, independent of visibility layer. (so it allows to move such lights to hidden layer, not influencing anything).
2006-12-08 09:40:44 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
shi->u= -u;
shi->v= -v;
shi->xs= x;
shi->ys= y;
Baking now does material texture anti-aliasing if the OSA button is enabled, result of that is mainly visible for image textures.
2008-03-12 20:22:05 +00:00
shade_input_set_uv(shi);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
shade_input_set_normals(shi);
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
/* no normal flip */
if(shi->flippednor)
shade_input_flip_normals(shi);
Fix for bug #8299: baking from selected to active with ray transparency gave wrong results. Problem was the view vector was setup strange. While it is really not well defined, it is now set as the negation of the normal, so it is looking right at the surface, gives more predictable.
2008-02-19 17:51:58 +00:00
/* set up view vector to look right at the surface (note that the normal
* is negated in the renderer so it does not need to be done here) */
shi->view[0]= shi->vn[0];
shi->view[1]= shi->vn[1];
shi->view[2]= shi->vn[2];
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
}
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
static void bake_shade(void *handle, Object *ob, ShadeInput *shi, int quad, int x, int y, float u, float v, float *tvn, float *ttang)
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
{
BakeShade *bs= handle;
ShadeSample *ssamp= &bs->ssamp;
ShadeResult shr;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
VlakRen *vlr= shi->vlr;
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
/* init material vars */
memcpy(&shi->r, &shi->mat->r, 23*sizeof(float)); // note, keep this synced with render_types.h
shi->har= shi->mat->har;
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
if(bs->type==RE_BAKE_AO) {
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
ambient_occlusion(shi);
Added an option to bake AO normalized, without material settings taken into account. The normalized result can often be easier to integrate into different materials, e.g. for games.
2008-02-19 16:16:16 +00:00
* Displacement baking wasnt working with negative distances. * Added Normalize option for diplacement so everything in the 'Dist' range is mapped 0-1 * Increased the maximum Dist and Bias to1000.0 (was 10.0) * Added python utility function in BPyRender.py - bakeToPlane(...), to automate heightmap, normalmap generation for Crystalspace.
2008-02-22 14:27:46 +00:00
if(R.r.bake_flag & R_BAKE_NORMALIZE)
Added an option to bake AO normalized, without material settings taken into account. The normalized result can often be easier to integrate into different materials, e.g. for games.
2008-02-19 16:16:16 +00:00
VECCOPY(shr.combined, shi->ao)
else
ambient_occlusion_to_diffuse(shi, shr.combined);
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
}
else {
Added shadow baking
2008-06-05 22:07:59 +00:00
if (bs->type==RE_BAKE_SHADOW) /* Why do shadows set the color anyhow?, ignore material color for baking */
shi->r = shi->g = shi->b = 1.0f;
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
shade_input_set_shade_texco(shi);
Added shadow baking
2008-06-05 22:07:59 +00:00
if(!ELEM3(bs->type, RE_BAKE_NORMALS, RE_BAKE_TEXTURE, RE_BAKE_SHADOW))
Fix for bug #7863: AO made normal map baking unnecessarily slow. Fix for bug #8226: approximate AO doesn't bake.
2008-02-13 13:25:19 +00:00
shade_samples_do_AO(ssamp);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
if(shi->mat->nodetree && shi->mat->use_nodes) {
ntreeShaderExecTree(shi->mat->nodetree, shi, &shr);
shi->mat= vlr->mat; /* shi->mat is being set in nodetree */
}
else
shade_material_loop(shi, &shr);
if(bs->type==RE_BAKE_NORMALS) {
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
float nor[3];
VECCOPY(nor, shi->vn);
if(R.r.bake_normal_space == R_BAKE_SPACE_CAMERA);
else if(R.r.bake_normal_space == R_BAKE_SPACE_TANGENT) {
float mat[3][3], imat[3][3];
/* bitangent */
if(tvn && ttang) {
VECCOPY(mat[0], ttang);
Crossf(mat[1], tvn, ttang);
VECCOPY(mat[2], tvn);
}
else {
Normal map tangents are now not always averaged at vertices anymore, but only when the UV's are connected. That fixes some artifacts when baking and using tangent space normal maps. It does mean increased memory usage because it now stores 4 tangents per face like UV's, and increased processing time, but there's no simple way around that.
2008-03-07 15:38:56 +00:00
VECCOPY(mat[0], shi->nmaptang);
Crossf(mat[1], shi->vn, shi->nmaptang);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
VECCOPY(mat[2], shi->vn);
}
Mat3Inv(imat, mat);
Mat3MulVecfl(imat, nor);
}
else if(R.r.bake_normal_space == R_BAKE_SPACE_OBJECT)
Mat4Mul3Vecfl(ob->imat, nor); /* ob->imat includes viewinv! */
else if(R.r.bake_normal_space == R_BAKE_SPACE_WORLD)
Mat4Mul3Vecfl(R.viewinv, nor);
Normalize(nor); /* in case object has scaling */
shr.combined[0]= nor[0]/2.0f + 0.5f;
shr.combined[1]= 0.5f - nor[1]/2.0f;
shr.combined[2]= nor[2]/2.0f + 0.5f;
Render Baking: - Bugfix: vertex normals were still flipped around, to match viewpoint rendering. - New: option to bake a texture+material only
2006-11-19 18:44:54 +00:00
}
else if(bs->type==RE_BAKE_TEXTURE) {
Bugfix #5444 Additions in passrendering broke Baking a bit; what did not work was having a full render + AO. Was missing a new flag initialize...
2006-12-23 12:58:25 +00:00
shr.combined[0]= shi->r;
shr.combined[1]= shi->g;
shr.combined[2]= shi->b;
masked/alpha baking commit was missing a null check for imbuf, made textures also bake alpha
2008-03-09 18:06:48 +00:00
shr.alpha = shi->alpha;
Render Baking: - Bugfix: vertex normals were still flipped around, to match viewpoint rendering. - New: option to bake a texture+material only
2006-11-19 18:44:54 +00:00
}
Added shadow baking
2008-06-05 22:07:59 +00:00
else if(bs->type==RE_BAKE_SHADOW) {
VECCOPY(shr.combined, shr.shad);
shr.alpha = shi->alpha;
}
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
}
Fix for bug #7936: render baking selected to active now has a Bias value that is an offset along the normal when looking for the nearest face, which allows baking faces further away, e.g. an ID badge onto a shirt. Also fixes a bug baking to float images, for things other than displacement it didn't work sometimes, and a memory leak in the extend filter.
2008-02-18 18:14:19 +00:00
if(bs->rect_float) {
float *col= bs->rect_float + 4*(bs->rectx*y + x);
VECCOPY(col, shr.combined);
masked/alpha baking commit was missing a null check for imbuf, made textures also bake alpha
2008-03-09 18:06:48 +00:00
if (bs->type==RE_BAKE_ALL || bs->type==RE_BAKE_TEXTURE) {
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
col[3]= shr.alpha;
} else {
col[3]= 1.0;
}
Fix for bug #7936: render baking selected to active now has a Bias value that is an offset along the normal when looking for the nearest face, which allows baking faces further away, e.g. an ID badge onto a shirt. Also fixes a bug baking to float images, for things other than displacement it didn't work sometimes, and a memory leak in the extend filter.
2008-02-18 18:14:19 +00:00
}
else {
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
char *col= (char *)(bs->rect + bs->rectx*y + x);
Bugfix #5444 Additions in passrendering broke Baking a bit; what did not work was having a full render + AO. Was missing a new flag initialize...
2006-12-23 12:58:25 +00:00
col[0]= FTOCHAR(shr.combined[0]);
col[1]= FTOCHAR(shr.combined[1]);
col[2]= FTOCHAR(shr.combined[2]);
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
masked/alpha baking commit was missing a null check for imbuf, made textures also bake alpha
2008-03-09 18:06:48 +00:00
if (bs->type==RE_BAKE_ALL || bs->type==RE_BAKE_TEXTURE) {
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
col[3]= FTOCHAR(shr.alpha);
} else {
col[3]= 255;
}
}
if (bs->rect_mask) {
bs->rect_mask[bs->rectx*y + x] = BAKE_MASK_BAKED;
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
}
}
Fix for bug #7936: render baking selected to active now has a Bias value that is an offset along the normal when looking for the nearest face, which allows baking faces further away, e.g. an ID badge onto a shirt. Also fixes a bug baking to float images, for things other than displacement it didn't work sometimes, and a memory leak in the extend filter.
2008-02-18 18:14:19 +00:00
static void bake_displacement(void *handle, ShadeInput *shi, float dist, int x, int y)
* Displacement map baking This is an extension on the work Brecht already did to implement normal map baking. I've updated the release notes page here with info and pics: http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-12-29 05:17:19 +00:00
{
BakeShade *bs= handle;
float disp;
* Displacement baking wasnt working with negative distances. * Added Normalize option for diplacement so everything in the 'Dist' range is mapped 0-1 * Increased the maximum Dist and Bias to1000.0 (was 10.0) * Added python utility function in BPyRender.py - bakeToPlane(...), to automate heightmap, normalmap generation for Crystalspace.
2008-02-22 14:27:46 +00:00
if(R.r.bake_flag & R_BAKE_NORMALIZE && R.r.bake_maxdist) {
disp = (dist+R.r.bake_maxdist) / (R.r.bake_maxdist*2); /* alter the range from [-bake_maxdist, bake_maxdist] to [0, 1]*/
} else {
disp = 0.5 + dist; /* alter the range from [-0.5,0.5] to [0,1]*/
}
* Displacement map baking This is an extension on the work Brecht already did to implement normal map baking. I've updated the release notes page here with info and pics: http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-12-29 05:17:19 +00:00
if(bs->rect_float) {
float *col= bs->rect_float + 4*(bs->rectx*y + x);
col[0] = col[1] = col[2] = disp;
col[3]= 1.0f;
} else {
char *col= (char *)(bs->rect + bs->rectx*y + x);
col[0]= FTOCHAR(disp);
col[1]= FTOCHAR(disp);
col[2]= FTOCHAR(disp);
col[3]= 255;
}
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
if (bs->rect_mask) {
bs->rect_mask[bs->rectx*y + x] = BAKE_MASK_BAKED;
}
* Displacement map baking This is an extension on the work Brecht already did to implement normal map baking. I've updated the release notes page here with info and pics: http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-12-29 05:17:19 +00:00
}
Fix for error in approximate AO in last commit, made it really slow. Also, duplis are now taking into account, the proper way to exclude them is to set the material to be not traceable. Removed an unnecessary pointer from the VlakRen struct to save some memory, not really that significant, but still, saves 70 mb for 10 million faces.
2008-01-21 23:17:19 +00:00
static int bake_check_intersect(Isect *is, int ob, RayFace *face)
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
{
BakeShade *bs = (BakeShade*)is->userdata;
/* no direction checking for now, doesn't always improve the result
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
* (INPR(shi->facenor, bs->dir) > 0.0f); */
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
Fix for error in approximate AO in last commit, made it really slow. Also, duplis are now taking into account, the proper way to exclude them is to set the material to be not traceable. Removed an unnecessary pointer from the VlakRen struct to save some memory, not really that significant, but still, saves 70 mb for 10 million faces.
2008-01-21 23:17:19 +00:00
return (R.objectinstance[ob].obr->ob != bs->actob);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
}
Fix for bug #7936: render baking selected to active now has a Bias value that is an offset along the normal when looking for the nearest face, which allows baking faces further away, e.g. an ID badge onto a shirt. Also fixes a bug baking to float images, for things other than displacement it didn't work sometimes, and a memory leak in the extend filter.
2008-02-18 18:14:19 +00:00
static int bake_intersect_tree(RayTree* raytree, Isect* isect, float *start, float *dir, float sign, float *hitco, float *dist)
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
{
float maxdist;
int hit;
/* might be useful to make a user setting for maxsize*/
if(R.r.bake_maxdist > 0.0f)
maxdist= R.r.bake_maxdist;
else
Fix for bug #7936: render baking selected to active now has a Bias value that is an offset along the normal when looking for the nearest face, which allows baking faces further away, e.g. an ID badge onto a shirt. Also fixes a bug baking to float images, for things other than displacement it didn't work sometimes, and a memory leak in the extend filter.
2008-02-18 18:14:19 +00:00
maxdist= RE_ray_tree_max_size(R.raytree) + R.r.bake_biasdist;
VECADDFAC(isect->start, start, dir, -R.r.bake_biasdist);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
isect->end[0] = isect->start[0] + dir[0]*maxdist*sign;
isect->end[1] = isect->start[1] + dir[1]*maxdist*sign;
isect->end[2] = isect->start[2] + dir[2]*maxdist*sign;
hit = RE_ray_tree_intersect_check(R.raytree, isect, bake_check_intersect);
if(hit) {
hitco[0] = isect->start[0] + isect->labda*isect->vec[0];
hitco[1] = isect->start[1] + isect->labda*isect->vec[1];
hitco[2] = isect->start[2] + isect->labda*isect->vec[2];
Fix for bug #7936: render baking selected to active now has a Bias value that is an offset along the normal when looking for the nearest face, which allows baking faces further away, e.g. an ID badge onto a shirt. Also fixes a bug baking to float images, for things other than displacement it didn't work sometimes, and a memory leak in the extend filter.
2008-02-18 18:14:19 +00:00
*dist= VecLenf(start, hitco);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
}
return hit;
}
Baking now does material texture anti-aliasing if the OSA button is enabled, result of that is mainly visible for image textures.
2008-03-12 20:22:05 +00:00
static void bake_set_vlr_dxyco(BakeShade *bs, float *uv1, float *uv2, float *uv3)
{
VlakRen *vlr= bs->vlr;
float A, d1, d2, d3, *v1, *v2, *v3;
if(bs->quad) {
v1= vlr->v1->co;
v2= vlr->v3->co;
v3= vlr->v4->co;
}
else {
v1= vlr->v1->co;
v2= vlr->v2->co;
v3= vlr->v3->co;
}
/* formula derived from barycentric coordinates:
* (uvArea1*v1 + uvArea2*v2 + uvArea3*v3)/uvArea
* then taking u and v partial derivatives to get dxco and dyco */
A= (uv2[0] - uv1[0])*(uv3[1] - uv1[1]) - (uv3[0] - uv1[0])*(uv2[1] - uv1[1]);
if(fabs(A) > FLT_EPSILON) {
A= 0.5f/A;
d1= uv2[1] - uv3[1];
d2= uv3[1] - uv1[1];
d3= uv1[1] - uv2[1];
bs->dxco[0]= (v1[0]*d1 + v2[0]*d2 + v3[0]*d3)*A;
bs->dxco[1]= (v1[1]*d1 + v2[1]*d2 + v3[1]*d3)*A;
bs->dxco[2]= (v1[2]*d1 + v2[2]*d2 + v3[2]*d3)*A;
d1= uv3[0] - uv2[0];
d2= uv1[0] - uv3[0];
d3= uv2[0] - uv1[0];
bs->dyco[0]= (v1[0]*d1 + v2[0]*d2 + v3[0]*d3)*A;
bs->dyco[1]= (v1[1]*d1 + v2[1]*d2 + v3[1]*d3)*A;
bs->dyco[2]= (v1[2]*d1 + v2[2]*d2 + v3[2]*d3)*A;
}
else {
bs->dxco[0]= bs->dxco[1]= bs->dxco[2]= 0.0f;
bs->dyco[0]= bs->dyco[1]= bs->dyco[2]= 0.0f;
}
if(bs->obi->flag & R_TRANSFORMED) {
Mat3MulVecfl(bs->obi->nmat, bs->dxco);
Mat3MulVecfl(bs->obi->nmat, bs->dyco);
}
}
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
static void do_bake_shade(void *handle, int x, int y, float u, float v)
{
BakeShade *bs= handle;
VlakRen *vlr= bs->vlr;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
ObjectInstanceRen *obi= bs->obi;
Object *ob= obi->obr->ob;
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
float l, *v1, *v2, *v3, tvn[3], ttang[3];
int quad;
ShadeSample *ssamp= &bs->ssamp;
ShadeInput *shi= ssamp->shi;
/* fast threadsafe break test */
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(R.test_break(R.tbh))
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
return;
/* setup render coordinates */
if(bs->quad) {
v1= vlr->v1->co;
v2= vlr->v3->co;
v3= vlr->v4->co;
}
else {
v1= vlr->v1->co;
v2= vlr->v2->co;
v3= vlr->v3->co;
}
/* renderco */
l= 1.0f-u-v;
shi->co[0]= l*v3[0]+u*v1[0]+v*v2[0];
shi->co[1]= l*v3[1]+u*v1[1]+v*v2[1];
shi->co[2]= l*v3[2]+u*v1[2]+v*v2[2];
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
if(obi->flag & R_TRANSFORMED)
Mat4MulVecfl(obi->mat, shi->co);
Baking now does material texture anti-aliasing if the OSA button is enabled, result of that is mainly visible for image textures.
2008-03-12 20:22:05 +00:00
VECCOPY(shi->dxco, bs->dxco);
VECCOPY(shi->dyco, bs->dyco);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
quad= bs->quad;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
bake_set_shade_input(obi, vlr, shi, quad, 0, x, y, u, v);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
if(bs->type==RE_BAKE_NORMALS && R.r.bake_normal_space==R_BAKE_SPACE_TANGENT) {
shade_input_set_shade_texco(shi);
VECCOPY(tvn, shi->vn);
Normal map tangents are now not always averaged at vertices anymore, but only when the UV's are connected. That fixes some artifacts when baking and using tangent space normal maps. It does mean increased memory usage because it now stores 4 tangents per face like UV's, and increased processing time, but there's no simple way around that.
2008-03-07 15:38:56 +00:00
VECCOPY(ttang, shi->nmaptang);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
}
/* if we are doing selected to active baking, find point on other face */
if(bs->actob) {
Isect isec, minisec;
Fix for bug #7936: render baking selected to active now has a Bias value that is an offset along the normal when looking for the nearest face, which allows baking faces further away, e.g. an ID badge onto a shirt. Also fixes a bug baking to float images, for things other than displacement it didn't work sometimes, and a memory leak in the extend filter.
2008-02-18 18:14:19 +00:00
float co[3], minco[3], dist, mindist=0.0f;
Routine purge of compiler warnings * Most were uninitialised vars * Fixed whitespace in a few places * The change I made in rendercore.c -> do_bake_shade() was for an uninitialised var, but I hope it does't cause any rendering errors...
2007-12-30 23:27:35 +00:00
int hit, sign, dir=1;
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
/* intersect with ray going forward and backward*/
hit= 0;
memset(&minisec, 0, sizeof(minisec));
minco[0]= minco[1]= minco[2]= 0.0f;
Routine purge of compiler warnings * Most were uninitialised vars * Fixed whitespace in a few places * The change I made in rendercore.c -> do_bake_shade() was for an uninitialised var, but I hope it does't cause any rendering errors...
2007-12-30 23:27:35 +00:00
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
VECCOPY(bs->dir, shi->vn);
Routine purge of compiler warnings * Most were uninitialised vars * Fixed whitespace in a few places * The change I made in rendercore.c -> do_bake_shade() was for an uninitialised var, but I hope it does't cause any rendering errors...
2007-12-30 23:27:35 +00:00
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
for(sign=-1; sign<=1; sign+=2) {
memset(&isec, 0, sizeof(isec));
isec.mode= RE_RAY_MIRROR;
isec.faceorig= (RayFace*)vlr;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
isec.oborig= RAY_OBJECT_SET(&R, obi);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
isec.userdata= bs;
Routine purge of compiler warnings * Most were uninitialised vars * Fixed whitespace in a few places * The change I made in rendercore.c -> do_bake_shade() was for an uninitialised var, but I hope it does't cause any rendering errors...
2007-12-30 23:27:35 +00:00
Fix for bug #7936: render baking selected to active now has a Bias value that is an offset along the normal when looking for the nearest face, which allows baking faces further away, e.g. an ID badge onto a shirt. Also fixes a bug baking to float images, for things other than displacement it didn't work sometimes, and a memory leak in the extend filter.
2008-02-18 18:14:19 +00:00
if(bake_intersect_tree(R.raytree, &isec, shi->co, shi->vn, sign, co, &dist)) {
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
if(!hit || VecLenf(shi->co, co) < VecLenf(shi->co, minco)) {
minisec= isec;
Fix for bug #7936: render baking selected to active now has a Bias value that is an offset along the normal when looking for the nearest face, which allows baking faces further away, e.g. an ID badge onto a shirt. Also fixes a bug baking to float images, for things other than displacement it didn't work sometimes, and a memory leak in the extend filter.
2008-02-18 18:14:19 +00:00
mindist= dist;
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
VECCOPY(minco, co);
hit= 1;
* Displacement map baking This is an extension on the work Brecht already did to implement normal map baking. I've updated the release notes page here with info and pics: http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-12-29 05:17:19 +00:00
dir = sign;
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
}
}
}
Fix for displacement baking when it doesn't find a hit. Instead of writing 0.0 it now writes the neutral value of 0.5.
2008-04-28 14:54:24 +00:00
if (bs->type==RE_BAKE_DISPLACEMENT) {
if(hit)
bake_displacement(handle, shi, (dir==-1)? mindist:-mindist, x, y);
else
bake_displacement(handle, shi, 0.0f, x, y);
* Displacement map baking This is an extension on the work Brecht already did to implement normal map baking. I've updated the release notes page here with info and pics: http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-12-29 05:17:19 +00:00
return;
}
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
/* if hit, we shade from the new point, otherwise from point one starting face */
if(hit) {
vlr= (VlakRen*)minisec.face;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
obi= RAY_OBJECT_GET(&R, minisec.ob);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
quad= (minisec.isect == 2);
VECCOPY(shi->co, minco);
Routine purge of compiler warnings * Most were uninitialised vars * Fixed whitespace in a few places * The change I made in rendercore.c -> do_bake_shade() was for an uninitialised var, but I hope it does't cause any rendering errors...
2007-12-30 23:27:35 +00:00
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
u= -minisec.u;
v= -minisec.v;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
bake_set_shade_input(obi, vlr, shi, quad, 1, x, y, u, v);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
}
}
if(bs->type==RE_BAKE_NORMALS && R.r.bake_normal_space==R_BAKE_SPACE_TANGENT)
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
bake_shade(handle, ob, shi, quad, x, y, u, v, tvn, ttang);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
else
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
bake_shade(handle, ob, shi, quad, x, y, u, v, 0, 0);
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
}
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
static int get_next_bake_face(BakeShade *bs)
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
{
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
ObjectRen *obr;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
VlakRen *vlr;
Added names to UV and vertex color layers, and display them as a list. Added support for multiple UVs in the render engine. This also involved changing the way faces are stored, to allow data to be added optionally per 256 faces, same as the existing system for vertices. A UV layer can be specified in the Map Input panel and the Geometry node by name. Leaving this field blank will default to the active UV layer. Also added sharing of face selection and hiding between UV layers, and at the same time improved syncing with editmode selection and hiding. Still to do: - Multi UV support for fastshade. - Multires and NMesh preservation of multiple UV sets.
2006-12-21 13:47:27 +00:00
MTFace *tface;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
static int v= 0, vdone= 0;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
static ObjectInstanceRen *obi= NULL;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
if(bs==NULL) {
vlr= NULL;
v= vdone= 0;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
obi= R.instancetable.first;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
return 0;
}
BLI_lock_thread(LOCK_CUSTOM1);
Added names to UV and vertex color layers, and display them as a list. Added support for multiple UVs in the render engine. This also involved changing the way faces are stored, to allow data to be added optionally per 256 faces, same as the existing system for vertices. A UV layer can be specified in the Map Input panel and the Geometry node by name. Leaving this field blank will default to the active UV layer. Also added sharing of face selection and hiding between UV layers, and at the same time improved syncing with editmode selection and hiding. Still to do: - Multi UV support for fastshade. - Multires and NMesh preservation of multiple UV sets.
2006-12-21 13:47:27 +00:00
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
for(; obi; obi=obi->next, v=0) {
obr= obi->obr;
for(; v<obr->totvlak; v++) {
vlr= RE_findOrAddVlak(obr, v);
if((bs->actob && bs->actob == obr->ob) || (!bs->actob && (obr->ob->flag & SELECT))) {
Fix for bug #11650: removing uv layers in editmode did not work correct. Fix for bug #11661: bake to viewport uv layer instead of render uv layer.
2008-05-15 19:40:09 +00:00
tface= RE_vlakren_get_tface(obr, vlr, obr->bakemtface, NULL, 0);
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
if(tface && tface->tpage) {
Image *ima= tface->tpage;
ImBuf *ibuf= BKE_image_get_ibuf(ima, NULL);
float vec[4]= {0.0f, 0.0f, 0.0f, 0.0f};
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
if(ibuf==NULL)
continue;
if(ibuf->rect==NULL && ibuf->rect_float==NULL)
continue;
if(ibuf->rect_float && !(ibuf->channels==0 || ibuf->channels==4))
continue;
/* find the image for the first time? */
if(ima->id.flag & LIB_DOIT) {
ima->id.flag &= ~LIB_DOIT;
/* we either fill in float or char, this ensures things go fine */
if(ibuf->rect_float)
imb_freerectImBuf(ibuf);
/* clear image */
if(R.r.bake_flag & R_BAKE_CLEAR)
IMB_rectfill(ibuf, vec);
/* might be read by UI to set active image for display */
R.bakebuf= ima;
}
bs->obi= obi;
bs->vlr= vlr;
bs->vdone++; /* only for error message if nothing was rendered */
v++;
BLI_unlock_thread(LOCK_CUSTOM1);
return 1;
}
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
}
}
}
BLI_unlock_thread(LOCK_CUSTOM1);
return 0;
}
/* already have tested for tface and ima and zspan */
static void shade_tface(BakeShade *bs)
{
VlakRen *vlr= bs->vlr;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
ObjectInstanceRen *obi= bs->obi;
ObjectRen *obr= obi->obr;
Fix for bug #11650: removing uv layers in editmode did not work correct. Fix for bug #11661: bake to viewport uv layer instead of render uv layer.
2008-05-15 19:40:09 +00:00
MTFace *tface= RE_vlakren_get_tface(obr, vlr, obr->bakemtface, NULL, 0);
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
Image *ima= tface->tpage;
float vec[4][2];
int a, i1, i2, i3;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
/* check valid zspan */
if(ima!=bs->ima) {
bs->ima= ima;
The Big Image refactor! Please read: http://www.blender3d.org/cms/Imaging.834.0.html Or in short: - adding MultiLayer Image support - recoded entire Image API - better integration of movie/sequence Images Was a whole load of work... went down for a week to do this. So, will need a lot of testing! Will be in irc all evening.
2006-12-20 17:57:56 +00:00
bs->ibuf= BKE_image_get_ibuf(ima, NULL);
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
/* note, these calls only free/fill contents of zspan struct, not zspan itself */
zbuf_free_span(bs->zspan);
Threaded Shadowbuffers Creation =============================== - One shadowbuffer per thread. - Added more break tests in shadow buffer code. - Removed R.clipcrop global, solution is not nice yet, but at least threadsafe. - Fixed bug in strand render shadow buffer code.
2007-12-17 21:04:22 +00:00
zbuf_alloc_span(bs->zspan, bs->ibuf->x, bs->ibuf->y, R.clipcrop);
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
}
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
The Big Image refactor! Please read: http://www.blender3d.org/cms/Imaging.834.0.html Or in short: - adding MultiLayer Image support - recoded entire Image API - better integration of movie/sequence Images Was a whole load of work... went down for a week to do this. So, will need a lot of testing! Will be in irc all evening.
2006-12-20 17:57:56 +00:00
bs->rectx= bs->ibuf->x;
bs->recty= bs->ibuf->y;
bs->rect= bs->ibuf->rect;
bs->rect_float= bs->ibuf->rect_float;
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
bs->quad= 0;
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
if (bs->usemask) {
if (bs->ibuf->userdata==NULL) {
stupid mistake for setting the bake mask (only masked half the faces), also exposed a possible bug since the malloc and null checks were done without mutex locking.
2008-03-10 20:58:53 +00:00
BLI_lock_thread(LOCK_CUSTOM1);
if (bs->ibuf->userdata==NULL) { /* since the thread was locked, its possible another thread alloced the value */
bs->ibuf->userdata = (void *)MEM_callocN(sizeof(char)*bs->rectx*bs->recty, "BakeMask");
bs->rect_mask= (char *)bs->ibuf->userdata;
}
BLI_unlock_thread(LOCK_CUSTOM1);
} else {
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
bs->rect_mask= (char *)bs->ibuf->userdata;
}
}
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
/* get pixel level vertex coordinates */
for(a=0; a<4; a++) {
2.50: svn merge https://svn.blender.org/svnroot/bf-blender/trunk/blender -r19820:HEAD Notes: * Game and sequencer RNA, and sequencer header are now out of date a bit after changes in trunk. * I didn't know how to port these bugfixes, most likely they are not needed anymore. * Fix "duplicate strip" always increase the user count for ipo. * IPO pinning on sequencer strips was lost during Undo.
2009-06-08 20:08:19 +00:00
/* Note, workaround for pixel aligned UVs which are common and can screw up our intersection tests
* where a pixel gets inbetween 2 faces or the middle of a quad,
* camera aligned quads also have this problem but they are less common.
* Add a small offset to the UVs, fixes bug #18685 - Campbell */
vec[a][0]= tface->uv[a][0]*(float)bs->rectx - (0.5f + 0.001);
vec[a][1]= tface->uv[a][1]*(float)bs->recty - (0.5f + 0.002);
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
}
/* UV indices have to be corrected for possible quad->tria splits */
i1= 0; i2= 1; i3= 2;
vlr_set_uv_indices(vlr, &i1, &i2, &i3);
Baking now does material texture anti-aliasing if the OSA button is enabled, result of that is mainly visible for image textures.
2008-03-12 20:22:05 +00:00
bake_set_vlr_dxyco(bs, vec[i1], vec[i2], vec[i3]);
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
zspan_scanconvert(bs->zspan, bs, vec[i1], vec[i2], vec[i3], do_bake_shade);
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
if(vlr->v4) {
bs->quad= 1;
Baking now does material texture anti-aliasing if the OSA button is enabled, result of that is mainly visible for image textures.
2008-03-12 20:22:05 +00:00
bake_set_vlr_dxyco(bs, vec[0], vec[2], vec[3]);
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
zspan_scanconvert(bs->zspan, bs, vec[0], vec[2], vec[3], do_bake_shade);
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
}
}
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
static void *do_bake_thread(void *bs_v)
{
BakeShade *bs= bs_v;
while(get_next_bake_face(bs)) {
shade_tface(bs);
/* fast threadsafe break test */
2.5 Added WM Jobs manager - WM can manage threaded jobs for you; just provide a couple of components to get it work: - customdata, free callback for it - timer step, notifier code - start callback, update callback - Once started, each job runs an own timer, and will for every time step check necessary updates, or close the job when ready. - No drawing happens in jobs, that's for notifiers! - Every job stores an owner pointer, and based on this owner it will prevent multiple jobs to enter the stack. Instead it will re-use a running job, signal it to stop and allow caller to re-initialize it even. - Check new wm_jobs.c for more explanation. Jobs API is still under construction. Fun: BLI_addtail(&wm->jobs, steve); :) Put Node shader previews back using wmJobs - Preview calculating is now fully threaded (1 thread still) - Thanks to new event system + notifiers, you can see previews update even while dragging sliders! - Currently it only starts when you change a node setting. Warning: the thread render shares Node data, so don't delete nodes while it renders! This topic is on the todo to make safe. Also: - bug in region initialize (do_versions) showed channel list in node editor wrong. - flagged the channel list 'hidden' now, it was really in the way! This is for later to work on anyway. - recoded Render API callbacks so it gets handlers passed on, no globals to use anymore, remember? - previewrender code gets now so much nicer! Will remove a lot of stuff from code soon.
2009-01-22 14:59:49 +00:00
if(R.test_break(R.tbh))
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
break;
}
bs->ready= 1;
return NULL;
}
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
/* using object selection tags, the faces with UV maps get baked */
/* render should have been setup */
Bugfix #5277 Bake-render: Quad faces still didn't get handled properly, error visible for vertex color or UV textures. Also: added error meny when a Bake cannot work because there are no Images or no Images with buffers
2006-11-21 18:09:32 +00:00
/* returns 0 if nothing was handled */
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
int RE_bake_shade_all_selected(Render *re, int type, Object *actob)
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
{
Render Monster support: (part 1) Removed all limitations from render code for maximum threads. The only define for this now is in BLI_threads.h, and currently set to 8. Note that each thread renders an entire tile, and also allocates the buffers for the tiles, so; more threads might work better with smaller tiles. IMPORTANT: node system won't work yet with more than 2 threads! So, don't try material nodes or compositing with over 2 threads. That I'll commit later today. What does work (should work :) is AO and soft shadow now.
2006-11-29 17:01:09 +00:00
BakeShade handles[BLENDER_MAX_THREADS];
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
ListBase threads;
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
Image *ima;
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
int a, vdone=0, usemask=0;
Render Instancing ================= Big commit, but little user visible changes. - Dupliverts and duplifaces are now rendered as instances, instead of storing all of the geometry for each dupli, now an instance is created with a matrix transform refering to the source object. This should allow us to render tree leaves more memory efficient. - Radiosity and to some degree raytracing of such objects is not really efficient still. For radiosity this is fundamentally hard to solve, but raytracing an octree could be created for each object, but the current octree code with it's fixed size doesn't allow this efficiently. - The regression tests survived, but with I expect that some bugs will pop up .. hopefully not too many :). Implementation Notes ==================== - Dupligroups and linked meshes are not rendered as instances yet, since they can in fact be different due to various reasons, instancing of these types of duplis that are the same can be added for them at a later point. - Each ObjectRen now stores it's own database, instead of there being one big databases of faces, verts, .. . Which objects that are actually rendered are defined by the list of ObjectRenInstances, which all refer to an ObjectRen. - Homogeneous coordinatess and clipping is now not stored in vertices anymore, but instead computed on the fly. This couldn't work for instances. That does mean some extra computation has to be done, but memory lookups can be slow too, and this saves some memory. Overall I didn't find a significant speed impact. - OSA rendering for solid and ztransp now is different. Instead of e.g. going 8 times over the databases times and rendering the z-buffer, it now goes over the database once and renders each polygon 8 times. That was necessary to keep instances efficient, and can also give some performance improvement without instances. - There was already instancing support in the yafray export code, now it uses Blender's render instances for export. - UV and color layer storage in the render was a bit messy before, now should be easier to understand. - convertblender.c was reorganized somewhat. Regular render, speedvector and baking now use a single function to create the database, previously there was code duplicated for it. - Some of these changes were done with future multithreading of scene and shadow buffer creation in mind, though especially for scene creation much work remains to be done to make it threadsafe, since it also involves a lot of code from blenkernel, and there is an ugly conflict with the way dupli groups work here .. though in the render code itself it's almost there.
2007-12-15 20:41:45 +00:00
/* initialize render global */
R= *re;
R.bakebuf= NULL;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
/* initialize static vars */
get_next_bake_face(NULL);
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
/* do we need a mask? */
if (re->r.bake_filter && (re->r.bake_flag & R_BAKE_CLEAR)==0)
usemask = 1;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
/* baker uses this flag to detect if image was initialized */
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
for(ima= G.main->image.first; ima; ima= ima->id.next) {
ImBuf *ibuf= BKE_image_get_ibuf(ima, NULL);
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
ima->id.flag |= LIB_DOIT;
masked/alpha baking commit was missing a null check for imbuf, made textures also bake alpha
2008-03-09 18:06:48 +00:00
if (ibuf)
ibuf->userdata = NULL; /* use for masking if needed */
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
}
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
Render Monster support: (part 1) Removed all limitations from render code for maximum threads. The only define for this now is in BLI_threads.h, and currently set to 8. Note that each thread renders an entire tile, and also allocates the buffers for the tiles, so; more threads might work better with smaller tiles. IMPORTANT: node system won't work yet with more than 2 threads! So, don't try material nodes or compositing with over 2 threads. That I'll commit later today. What does work (should work :) is AO and soft shadow now.
2006-11-29 17:01:09 +00:00
BLI_init_threads(&threads, do_bake_thread, re->r.threads);
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
/* get the threads running */
Render Monster support: (part 1) Removed all limitations from render code for maximum threads. The only define for this now is in BLI_threads.h, and currently set to 8. Note that each thread renders an entire tile, and also allocates the buffers for the tiles, so; more threads might work better with smaller tiles. IMPORTANT: node system won't work yet with more than 2 threads! So, don't try material nodes or compositing with over 2 threads. That I'll commit later today. What does work (should work :) is AO and soft shadow now.
2006-11-29 17:01:09 +00:00
for(a=0; a<re->r.threads; a++) {
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
/* set defaults in handles */
memset(&handles[a], 0, sizeof(BakeShade));
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
handles[a].ssamp.shi[0].lay= re->scene->lay;
Added shadow baking
2008-06-05 22:07:59 +00:00
if (type==RE_BAKE_SHADOW) {
handles[a].ssamp.shi[0].passflag= SCE_PASS_SHADOW;
} else {
handles[a].ssamp.shi[0].passflag= SCE_PASS_COMBINED;
}
Bugfix #5444 Additions in passrendering broke Baking a bit; what did not work was having a full render + AO. Was missing a new flag initialize...
2006-12-23 12:58:25 +00:00
handles[a].ssamp.shi[0].combinedflag= ~(SCE_PASS_SPEC);
Next stage of RenderPipe refactor: now everything within the pixel was tackled. Resulting features: - render passes - new pass: Object Index, for masking - sub-sample alpha masks Docs: http://mediawiki.blender.org/index.php/BlenderDev/RenderPipeline http://www.blender3d.org/cms/Render_Passes.829.0.html http://www.blender3d.org/cms/New_Render_features.774.0.html Note that these changes might mean things to not render fully identical... For the next days a lot of testing is needed!
2006-12-05 16:43:01 +00:00
handles[a].ssamp.shi[0].thread= a;
Bugfix #5444 Additions in passrendering broke Baking a bit; what did not work was having a full render + AO. Was missing a new flag initialize...
2006-12-23 12:58:25 +00:00
handles[a].ssamp.tot= 1;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
handles[a].type= type;
Render Baking ============= A new "Selected to Active" option in the Bake panel, to (typically) bake a high poly object onto a low poly object. Code based on patch #7339 by Frank Richter (Crystal Space developer), thanks!. Normal Mapping ============== Camera, World, Object and Tangent space is now supported for baking, and for material textures. The "NMap TS" setting is replaced with a dropdown of the four choices in the image texture buttons. http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-11-28 22:21:12 +00:00
handles[a].actob= actob;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
handles[a].zspan= MEM_callocN(sizeof(ZSpan), "zspan for bake");
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
handles[a].usemask = usemask;
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
BLI_insert_thread(&threads, &handles[a]);
}
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
/* wait for everything to be done */
a= 0;
Render Monster support: (part 1) Removed all limitations from render code for maximum threads. The only define for this now is in BLI_threads.h, and currently set to 8. Note that each thread renders an entire tile, and also allocates the buffers for the tiles, so; more threads might work better with smaller tiles. IMPORTANT: node system won't work yet with more than 2 threads! So, don't try material nodes or compositing with over 2 threads. That I'll commit later today. What does work (should work :) is AO and soft shadow now.
2006-11-29 17:01:09 +00:00
while(a!=re->r.threads) {
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
PIL_sleep_ms(50);
Render Monster support: (part 1) Removed all limitations from render code for maximum threads. The only define for this now is in BLI_threads.h, and currently set to 8. Note that each thread renders an entire tile, and also allocates the buffers for the tiles, so; more threads might work better with smaller tiles. IMPORTANT: node system won't work yet with more than 2 threads! So, don't try material nodes or compositing with over 2 threads. That I'll commit later today. What does work (should work :) is AO and soft shadow now.
2006-11-29 17:01:09 +00:00
for(a=0; a<re->r.threads; a++)
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
if(handles[a].ready==0)
break;
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
}
* Displacement map baking This is an extension on the work Brecht already did to implement normal map baking. I've updated the release notes page here with info and pics: http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-12-29 05:17:19 +00:00
/* filter and refresh images */
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
for(ima= G.main->image.first; ima; ima= ima->id.next) {
Bake Render bug: check for which image to filter (bleed) used wrong flag, picking wrong images (or even ones without buffer, so crashing).
2006-11-28 18:28:23 +00:00
if((ima->id.flag & LIB_DOIT)==0) {
The Big Image refactor! Please read: http://www.blender3d.org/cms/Imaging.834.0.html Or in short: - adding MultiLayer Image support - recoded entire Image API - better integration of movie/sequence Images Was a whole load of work... went down for a week to do this. So, will need a lot of testing! Will be in irc all evening.
2006-12-20 17:57:56 +00:00
ImBuf *ibuf= BKE_image_get_ibuf(ima, NULL);
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
if (re->r.bake_filter) {
if (usemask) {
/* extend the mask +2 pixels from the image,
* this is so colors dont blend in from outside */
char *temprect;
for(a=0; a<re->r.bake_filter; a++)
bake_mask_filter_extend((char *)ibuf->userdata, ibuf->x, ibuf->y);
temprect = MEM_dupallocN(ibuf->userdata);
/* expand twice to clear this many pixels, so they blend back in */
bake_mask_filter_extend(temprect, ibuf->x, ibuf->y);
bake_mask_filter_extend(temprect, ibuf->x, ibuf->y);
/* clear all pixels in the margin*/
bake_mask_clear(ibuf, temprect, BAKE_MASK_MARGIN);
MEM_freeN(temprect);
}
for(a=0; a<re->r.bake_filter; a++) {
/*the mask, ibuf->userdata - can be null, in this case only zero alpha is used */
IMB_filter_extend(ibuf, (char *)ibuf->userdata);
}
if (ibuf->userdata) {
MEM_freeN(ibuf->userdata);
ibuf->userdata= NULL;
}
}
The Big Image refactor! Please read: http://www.blender3d.org/cms/Imaging.834.0.html Or in short: - adding MultiLayer Image support - recoded entire Image API - better integration of movie/sequence Images Was a whole load of work... went down for a week to do this. So, will need a lot of testing! Will be in irc all evening.
2006-12-20 17:57:56 +00:00
ibuf->userflags |= IB_BITMAPDIRTY;
* Displacement map baking This is an extension on the work Brecht already did to implement normal map baking. I've updated the release notes page here with info and pics: http://www.blender.org/development/current-projects/changes-since-244/render-baking/
2007-12-29 05:17:19 +00:00
if (ibuf->rect_float) IMB_rect_from_float(ibuf);
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
}
}
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
/* calculate return value */
Bugfix [#8196] sequencer crash when adding scene without camera added checks for a camera Also added an error ob bake if there are no active objects.
2008-02-17 20:44:07 +00:00
for(a=0; a<re->r.threads; a++) {
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
vdone+= handles[a].vdone;
zbuf_free_span(handles[a].zspan);
MEM_freeN(handles[a].zspan);
* Bakeing fix for apricot, stop pixels being extended for parts of the image that are not being baked (useful for multiple objects) * Added support for baking alpha, as well as extending alpha margins.
2008-03-08 21:35:38 +00:00
}
Render Baking upgrade: - Now baking itself is threaded too (like for render, max 2 cpus. Moving this to 4 cpus is on todo. Goes twice as fast! - fix: ESC from bake was broken... - other fix: toolbox menus didn't treat sublevel string lengths OK, truncating items like for Group library names.
2006-11-28 10:16:24 +00:00
BLI_end_threads(&threads);
Bugfix #5277 Bake-render: Quad faces still didn't get handled properly, error visible for vertex color or UV textures. Also: added error meny when a Bake cannot work because there are no Images or no Images with buffers
2006-11-21 18:09:32 +00:00
return vdone;
Long waited feature: Render Baking Here's the full release log with example file. http://www.blender3d.org/cms/Render_Baking.827.0.html For people who don't read docs; just press ALT+CTRL+B on a Mesh with texture faces! Todos: - maybe some filter options extra? - Make normal maps in Tangent space
2006-11-19 14:12:56 +00:00
}
Render Baking candy: while baking, the UV Image window shows the progress in rendered images, once per second. Also switches to other images. This happens threaded, so might need a good test :)
2006-11-22 18:45:34 +00:00
struct Image *RE_bake_shade_get_image(void)
{
return R.bakebuf;
}
Reference in New Issue Copy Permalink