Remove limit for number of textures when using CPU rendering,
so huge scenes from Mango could be rendered properly.
Re-arranged float and bytes textures for this, so now float
textures goes before byte. This is needed to make alignment
stuff easier for variable number of textures.
Also replace code-generated CPU kernel structure's members
with an array, which makes it easier to access textures
without adding bunch of cases in switch.
Should be no functional changes when rendering with GPU
or OpenCL devices.
Update samples displayign in the interface once in a whiel only
(currently once in 1 sec). This still keeps rendering interactive
enough for artists and it eliminates slowdown caused by constant
uploading sampels from GPU to host.
This also allowed to get rid of hack with storing render result
in render buffers, due to it's not so big overhead ow to create
render result when sample needs to be updated.
Tests made with different files shows that now render speed is
really close to oroginal speed from before we started hackign
into this stuff.
There still some issues with interactive update when rendering
several render layers, but that seems to be irrelivant to all
this sampels commtis, so woudl look into this as a separate
patch.
This fixes memory overflow caused by creating render result every time
RenderResult is creating when updating sample and not being freed until
tile is fully rendered.
Solved in probably not best way -- RenderResult is being stored in
RenderBuffers, so it's creating only once.
This solves memory issues, but while was looking into this issue
discovered dramatic slowdown caused by samples update in some files
from mango svn.
Solving this slowdown is becoming first priority from now on.
Was requested by Mango team to improve feedback during rendering.
Known issues:
- Updating of samples are accumulative, meaning that visually samples
would be dark in the beginning becoming brighter during progress.
- Could give some % of slowdown, so probably should be disabled in
background mode.
Still to come: update of samples when using CUDA and OpenCL.
selecting Box instead of Flat as projection, and then setting the blend value.
Blend 0.0 will not do any blending between box sides, higher values will give
a smooth transition between the sides
This is useful for mapping image texture without too obvious patterns in a
way that looks seamless, without the need for UV mapping. This works basically
the same as the mango node setup that was posted, just a bit faster:
http://mango.blender.org/production/blended_box/
- images that can't be loaded because of the limit are printed in the console.
- textures that can't be found show up as pink (so we know somethings wrong).
save memory during render and cache render results.
Implementation notes:
In the render engine API it's now possible to get the render result for
one render layer only, and retrieve the expected tile size in case save
buffers is used. This is needed because EXR expects tiles with particular
size and coordinates.
The EXR temporary files are now also separated per layer, since Cycles
can't give the full render result for all render layers, and EXR doesn't
support writing parts of tiles.
In Cycles internally the handling of render buffers and multi GPU
rendering in particular changed quite a bit, and could use a bit more
refactoring to make things more consistent and simple.
direct and indirect lighting differently. Rather than picking one light for each
point on the path, it now loops over all lights for direct lighting. For indirect
lighting it still picks a random light each time.
It gives control over the number of AA samples, and the number of Diffuse, Glossy,
Transmission, AO, Mesh Light, Background and Lamp samples for each AA sample.
This helps tuning render performance/noise and tends to give less noise for renders
dominated by direct lighting.
This sampling mode only works on the CPU, and still needs proper tile rendering
to show progress (will follow tommorrow or so), because each AA sample can be quite
slow now and so the delay between each update wil be too long.
The particle data is stored in a separate texture if any of the dupli objects uses particle info nodes in shaders. To map dupli objects onto particles the store an additional particle_index value, which is different from the simple dupli object index (only visible particles, also works for particle dupli groups mode).
Some simple use cases on the code.blender.org blog:
http://code.blender.org/index.php/2012/05/particle-info-node/
* More fixes for r41599, removed clouds and distorted noise textures and ported the Noise texture to OSL.
ToDo: Color output is still commented, needs a closer look.
* Some more fixes (comments, uninitialized variables)
* The random output had a range from -0.5 to 0.5, added an offset of 0.5 to bring it into the 0.0 / 1.0 range, which is in alignment with Shader Node Guidelines.
Patch by Agustin Benavidez, thanks!
pass index, and a random number unique to the instance of the object.
This can be useful to give some variation to a single material assigned to
multiple instances, either manually controlled through the object index, based
on the object location, or randomized for each instance.
http://wiki.blender.org/index.php/Doc:2.6/Manual/Render/Cycles/Nodes/More#Object_Info
* Multithreaded image loading, each thread can load a separate image.
* Better multithreading for multiple instanced meshes, different threads can now
build BVH's for different meshes, rather than all cooperating on the same mesh.
Especially noticeable for dynamic BVH building for the viewport, gave about
2x faster build on 8 core in fairly complex scene with many objects.
* The main thread waiting for worker threads can now also work itself, so
(num_cores + 1) threads will be working, this supposedly gives better
performance on some operating systems, but did not measure performance for
this very detailed yet.
For sample images see:
http://www.dalaifelinto.com/?p=399 (equisolid)
http://www.dalaifelinto.com/?p=389 (equidistant)
The 'use_panorama' option is now part of a new Camera type: 'Panorama'.
Created two other panorama cameras:
- Equisolid: most of lens in the market simulate this lens - e.g. Nikon, Canon, ...)
this works as a real lens up to an extent. The final result takes the
sensor dimensions into account also.
.:. to simulate a Nikon DX2S with a 10.5mm lens do:
sensor: 23.7 x 15.7
fisheye lens: 10.5
fisheye fov: 180
render dimensions: 4288 x 2848
- Equidistant: this is not a real lens model. Although the old equidistant lens simulate
this lens. The result is always as a circular fisheye that takes the whole sensor
(in other words, it doesn't take the sensor into consideration).
This is perfect for fulldomes ;)
For the UI we have 10 to 360 as soft values and 10 to 3600 as hard values (because we can).
Reference material:
http://www.hdrlabs.com/tutorials/downloads_files/HDRI%20for%20CGI.pdfhttp://www.bobatkins.com/photography/technical/field_of_view.html
Note, this is not a real simulation of the light path through the lens.
The ideal solution would be this:
https://graphics.stanford.edu/wikis/cs348b-11/Assignment3http://www.graphics.stanford.edu/papers/camera/
Thanks Brecht for the fix, suggestions and code review.
Kudos for the dome community for keeping me stimulated on the topic since 2009 ;)
Patch partly implemented during lab time at VisGraf, IMPA - Rio de Janeiro.
Most of the changes are related to adding support for motion data throughout
the code. There's some code for actual camera/object motion blur raytracing
but it's unfinished (it badly slows down the raytracing kernel even when the
option is turned off), so that code it disabled still.
Motion vector export from Blender tries to avoid computing derived meshes
when the mesh does not have a deforming modifier, and it also won't store
motion vectors for every vertex if only the object or camera is moving.
=== BVH build time optimizations ===
* BVH building was multithreaded. Not all building is multithreaded, packing
and the initial bounding/splitting is still single threaded, but recursive
splitting is, which was the main bottleneck.
* Object splitting now uses binning rather than sorting of all elements, using
code from the Embree raytracer from Intel.
http://software.intel.com/en-us/articles/embree-photo-realistic-ray-tracing-kernels/
* Other small changes to avoid allocations, pack memory more tightly, avoid
some unnecessary operations, ...
These optimizations do not work yet when Spatial Splits are enabled, for that
more work is needed. There's also other optimizations still needed, in
particular for the case of many low poly objects, the packing step and node
memory allocation.
BVH raytracing time should remain about the same, but BVH build time should be
significantly reduced, test here show speedup of about 5x to 10x on a dual core
and 5x to 25x on an 8-core machine, depending on the scene.
=== Threads ===
Centralized task scheduler for multithreading, which is basically the
CPU device threading code wrapped into something reusable.
Basic idea is that there is a single TaskScheduler that keeps a pool of threads,
one for each core. Other places in the code can then create a TaskPool that they
can drop Tasks in to be executed by the scheduler, and wait for them to complete
or cancel them early.
=== Normal ====
Added a Normal output to the texture coordinate node. This currently
gives the object space normal, which is the same under object animation.
In the future this might become a "generated" normal so it's also stable for
deforming objects, but for now it's already useful for non-deforming objects.
=== Render Layers ===
Per render layer Samples control, leaving it to 0 will use the common scene
setting.
Environment pass will now render environment even if film is set to transparent.
Exclude Layers" added. Scene layers (all object that influence the render,
directly or indirectly) are shared between all render layers. However sometimes
it's useful to leave out some object influence for a particular render layer.
That's what this option allows you to do.
=== Filter Glossy ===
When using a value higher than 0.0, this will blur glossy reflections after
blurry bounces, to reduce noise at the cost of accuracy. 1.0 is a good
starting value to tweak.
Some light paths have a low probability of being found while contributing much
light to the pixel. As a result these light paths will be found in some pixels
and not in others, causing fireflies. An example of such a difficult path might
be a small light that is causing a small specular highlight on a sharp glossy
material, which we are seeing through a rough glossy material. With path tracing
it is difficult to find the specular highlight, but if we increase the roughness
on the material the highlight gets bigger and softer, and so easier to find.
Often this blurring will be hardly noticeable, because we are seeing it through
a blurry material anyway, but there are also cases where this will lead to a
loss of detail in lighting.
