The Asset Browser (as opposed to other Filebrowser) is restricted to
either FILE_HORIZONTALDISPLAY / FILE_IMGDISPLAY, as noted in https://
projects.blender.org/blender/blender/issues/103563#issuecomment-82835
the UI isnt polished enough for FILE_VERTICALDISPLAY it seems. However,
resetting to default would leave us with exactly that.
To resolve, control the default dynamically via an
`RNA_def_property_enum_default_func` (and set it to FILE_IMGDISPLAY for
the Asset Browser).
Pull Request: https://projects.blender.org/blender/blender/pulls/144917
Consider the part of the view name before the dash to be the prefix that
we expect the look to have. So that both "AgX" and "AgX - HDR" match
looks starting with "AgX -".
Pull Request; https://projects.blender.org/blender/blender/pulls/142758
Almost all settings were duplicated between BakeData and RenderData.
The only missing field was the bake type, which is stored as a custom
property in Cycles.
This change:
- Removes unused bake_samples and bake_biasdist.
- Migrates settings like bake_margin to BakeData.
- Switches multires baker to use bake_margin.
- Introduces bake type in the BakeData, the same way how it was
defined in RenderData::bake_mode.
Pull Request: https://projects.blender.org/blender/blender/pulls/144984
This change makes it so baking displacement to the high-resolution mesh
(Use Low Resolution Mesh option is OFF, displacement is calculated
between top multi-resolution level and subdivided viewport level mesh)
uses texture UVs and tangent space from the high-res mesh.
This matches intended use-case when object baked with such configuration
have subdivision surface applied to them bringing overall resolution to
the same level as the highest multi-resolution level.
The issue was simple to see when baking high-res displacement for an
object which uses "Keep Corners, Junctions" UV smoothing.
The unfortunate aspect is increased memory usage due to calculation of
the grid index and grid UV mapping, but it is not too bad (12 bytes per
loop, so is like 48Mb per million face). Feels like there is a way to
optimize it by utilizing knowledge that high-res mesh faces are created
in a specific order, but also feels it is not that important at this
moment.
Pull Request: https://projects.blender.org/blender/blender/pulls/144935
The main idea is to switch Bake from Multires from legacy DerivedMesh
to Subdiv. On the development side of things this change removes a lot
of code, also making it easier easier to rework CustomData and related
topics, without being pulled down by the DerivedMesh.
On the user level switch to Subdiv means:
- Much more closer handling of the multi-resolution data: the derived
mesh code was close, but not exactly the same when it comes to the
final look of mesh.
Other than less obvious cases (like old DerivedMesh approach doing
recursive subdivision instead of pushing subdivided vertices on the
limit surface) there are more obvious ones like difference in edge
creases, and non-supported vertex creases by the DerivedMesh.
- UV interpolation is done correctly now when baking to non-base level
(baking to multi-resolution level >= 1).
Previously in this case the old derived mesh interpolation was used
to interpolate face-varying data, which gives different results from
the OpenSubdiv interpolation.
- Ngon faces are properly supported now.
A possible remaining issue is the fact that getting normal from CCG
always uses smooth interpolation. Based on the code it always has been
the case, so while it is something to look into it might be considered
a separate topic to dig into.
It matches the behavior of IMB_rectfill_area() with opaque alpha now.
This is important as the function is used in baking, where the neutral
0.5 is expected to be mapped to 128.
The only annoying part is that for the external MDISPS the mesh
needs to be cast to non-const.
Arguably it is semantically correct behavior as the details of
how external data is read is kind of implementation details.
- Return evaluated points as function return value.
Used in cases when only limit or final point is requested, without
partial derivatives.
- Re-order function argument list and use default argument values to
simplify usage in the simple/typical cases.
No functional changes expected.
This PR adds recycling of descriptor pools. Currently descriptor pools
are discarded when full or context is flushed. This PR allows
descriptor pools to be discarded for reuse.
It is also more conservative and only discard
Descriptor pools when they are full or fragmented.
When using the Vulkan backend a small amount of descriptor memory can leak. Even
when we clean up all resources, drivers can still keep data around on the
GPU. Eventually this can lead to out of memory issues depending on how
the GPU driver actually manages descriptor sets.
When the descriptor sets of the descriptor pool aren't used anymore
the VKDiscardPool will recycle the pools back to its original VKDescriptorPools.
It needs to be the same instance as descriptor pools/sets are owned by
a single thread.
Pull Request: https://projects.blender.org/blender/blender/pulls/144992
get_texture_colorspace_name accessed byte_buffer instead of
float_buffer; introduced in 18110744a2. In practice VSE seems to
always fill colorspaces, so does not cause a very visible bug.
Instead of matching regions, free all popups using
UI_popup_handlers_remove_all which also calls the remove callback
that frees the uiPopupBlockHandle.
Deleting or changing the screen with a popup (error report in this case)
wasn't removing the popup, it would attempt to display using it's
region which had been freed when changing the screen.
Regression in [0] an unsaved file would be an empty string which could
be equal to "libname" causing them to be considered equal, resulting
in an error thatthe file was linking in itself.
Resolve by only performing this check on saved files.
fa29d28d73
This PR adds HDR support for Windows for `VK_COLOR_SPACE_EXTENDED_SRGB_LINEAR_EXT`
on `VK_FORMAT_R16G16B16A16_SFLOAT` swapchains .
For nonlinear surface formats (sRGB and extended sRGB) the back buffer is blit into the swapchain,
When VK_COLOR_SPACE_EXTENDED_SRGB_LINEAR_EXT is used as surface format a compute shader
is used to flip and invert the gamma.
SDR white level is updated from a few window event changes, but actually
none of them immediately respond to SDR white level changes in the system.
That requires using the WinRT API, which we don't do so far.
Current limitations:
- Intel GPU support
- Dual GPU support
In the future we may add controls inside Blender for absolute HDR nits,
across different platforms. But this makes behavior closer to macOS.
See !144565 for details
Co-authored-by: Brecht Van Lommel <brecht@blender.org>
Pull Request: https://projects.blender.org/blender/blender/pulls/144717
Pool of VolumeInfo was never reset. Detected when stress testing
the vulkan backend using tree_creature.blend where the world
volume was converted to a volume object.
Pull Request: https://projects.blender.org/blender/blender/pulls/144967
When disabling cursor constraint NumInput::idx_max was set to -1.
This looks to be the behavior since 2.4x, however more recently
the modulo of `idx_max + 1` was used to cycle the axis when pressing
tab (crashing when `idx_max` is -1).
Resolve by setting `NumInput::idx_max` to the number of axes since
this is the state when transform is initialized.
Only create an outliner tree element for all Action slots in specific
modes of the outliner. The "object-hierarchical" modes (Scene, View
Layer) show the Action underneath each user of that Action; in these
cases, only the assigned slot is shown now.
Previously all slots would be shown, in all modes of the Outliner. This
meant quadratic complexity, as each Action is included in the tree for
each of its users, and then each of those tree items would show all
slots. This means the number of tree items for the slots was multiplied
by the number of users.
Pull Request: https://projects.blender.org/blender/blender/pulls/144864
In the Node Editor itself, you cannot edit a linked Nodetree (add nodes,
drag from a socket etc., so basically no actions allowed).
In the Properties Editor (Material tab) however, where we display the
Nodetree differently (via `uiTemplateNodeView`) and while editing
properties themselves was already disabled, you could still edit node
**links** (thus **adding** nodes to the tree).
This would all be lost on file reload though, so better prevent this.
To resolve, just disable the socket / NodeLink button when looking at a
linked Nodetree.
NOTE: the "not in a linked nodetree" disabled hint is not showing yet,
something I intend to fix in an upcoming PR (it is also not showing this
for properties generally in the node editor).
Pull Request: https://projects.blender.org/blender/blender/pulls/144930
* Bundled ICC profiles for display spaces supported by Blender, and
embed them in the image file when saving.
* Verified to work for PNG, TIFF, JPEG and WebP, but not all file
formats support this.
* No ICC profile is written for sRGB currently. It would be a matter
of adding an icc file, however this may be a breaking change for
some use cases.
* Fix save as render of EXR files not properly changing the image
colorspace to match.
Uses CC0 licensed ICC files from the Compact ICC Profiles project.
This does not include support for saving HDR images. While there exist
ICC profiles for PQ, they are not well supported and the preferred method
for HDR is to write CICP tags. However OpenImageIO support for this is
still under development.
Ref #144911
Pull Request: https://projects.blender.org/blender/blender/pulls/144565
These should at least show the correct colors and gamut now, following the
color picking space in the OpenColorIO config.
However the ACES config use sRGB as the color_picking role, which means
that the color picker only shows a limited a gamut. There is apparently is
not agreement on the interpretation of this role, so we might have to stop
using it.
Additionally, if you set the color_picking role to something like ACEScg +
sRGB transfer function, the widgets do show all the colors in the gamut,
however the gamut is probably too large to be practical for most uses. So
we need some UI solution for this regardless.
Ref #144911
Pull Request: https://projects.blender.org/blender/blender/pulls/144565
This was a temporary solution until we had proper HDR displays.
* Auto detect from the display + view transform if they are HDR, and
enable it automatically. This is based on encoding hdr-video in the config.
* If a HDR transform is selected and there is no HDR display support, an
info message will be shown in the color management panel.
* It is now possible to view HDR images in the image editor, without
needing to use "View as Render".
* There is no versioning to switch to a HDR display, because that also
affects image saving. So users will have to manually select the
"Rec.2100 PQ" display to see HDR colors again.
Ref #144911
Pull Request: https://projects.blender.org/blender/blender/pulls/144565
* Change meaning of display device to be the device that we are trying
to emulate, not the monitor configuration.
* The display transform is now:
* User specified view and display transform
* Clamp to 0 to limit colors to gamut
* Inverse untonemapped display transform
* Convert to extended sRGB
* When using the display space in the color management API, it now needs to
be specified if it's for drawing (with emulation) or file output or color
inspection (without).
Like HDR, this only works on macOS and Linux + Wayland + Vulkan currently.
Support for Windows/Vulkan is under development in #144717.
Ref #144911
Pull Request: https://projects.blender.org/blender/blender/pulls/144565
For display and image saving, we need to know certain properties of displays
and view transforms that OpenColorIO does not explicitly provide. In the
upcoming OpenColorIO 2.5 there will be an interop ID for color spaces that
adds more information. This moves towards that design, with heuristics for
configs that don't have the interop IDs.
* Try to use the first colorspace alias as the interop ID. This is used in
some existing configs, and now the Blender config as well.
* Improve ACES 2.0 config compatibility by interpreting some aliases like
srgb_display as interop IDs.
* Detect is_srgb and is_wide_gamut through display colorspace interop ID.
These are now per view rather than per display.
* Detect is_extended where we can't clamp to 0..1 through interop ID.
* Detect untonemapped view as config wide default view transform for converting
between reference and display space, if "Standard" or "Un-tone-mapped" can
not be found.
* Detect which display and view transform combination is HDR by checking for
"hdr-video" encoding on the display colorspace in the OpenColorIO config.
Ref #144911
Pull Request: https://projects.blender.org/blender/blender/pulls/144565
* PROP_COLOR_GAMMA is sRGB, not display space
* Hex colors are always sRGB
* Image byte buffers are in byte_buffer.colorspace
Fixes for sequencer text, image painting, render stamp and tooltips.
The default display space is sRGB, so this change will not be noticed
in most files.
Ref #144911
Pull Request: https://projects.blender.org/blender/blender/pulls/144565
DEG_add_collision_relations was only adding relationships for objects in
the brush collection, however dynamic paint also updates parent objects
and armature poses which weren't accounted for. Failing to include these
relations meant the evaluating the depsgraph could evaluate a pose
object from multiple threads at once - causing a crash.
Resolve by sharing logic between DEG_add_collision_relations &
BKE_object_modifier_update_subframe so depsgraph relationships match
objects the dynamic-paint modifier updates as part of its evaluation.
Ref !144844
Improve the documentation for the Bone coordinate spaces and these struct fields:
- `head` & `tail`
- `bone_mat`
- `arm_head` / `arm_tail`
- `dist`
No functional changes.
Pull Request: https://projects.blender.org/blender/blender/pulls/143729
