Since recent change in code of edge deduplication there is bunch of
reports about regressions whose were not found by tests. This commit
add some of recently explored cases to test coverage, right from
#147961. Not sure how to cover other know case from #147797 due to
use of USD file as source.
Pull Request: https://projects.blender.org/blender/blender/pulls/148237
Mesh invariants imply that edges and faces must be unique, so things
like reversed edges or duplicate faces with equal vertices are invalid.
For this reason, every time we generate new elements we have to ensure
that all new elements are unique between each other and already existing
elements. The recent refactor (ea875f6f32) introduced a new
algorithm to generate new mesh elements, and deduplication of new
elements was also a part of it. The problem is that the deduplication
only guaranteed that the original elements and new elements don't
overlap; deduplication between each new elements is not complete.
To solve the problem both new triangles and new edges have to be
deduplicated, even if there is no duplicates. Just to know this we have
to build a hash sets.
Triangle deduplication is a special part of the triangulation code,
but edges already handled elsewhere in the code base.
This refactor fixes this by replacing the original approach with one
which guarantees distinct faces and edges in the result.
Unfortunately, this fix increases runtime of the node 10x for a simple
cube with 500-vertex sides. It should be possible to make the
performance better again, but that requires more work.
Other work had to be done to enable this, so this depends on:
- [x] 157e7e0351
- [x] fa8574b80b
Co-authored-by: Hans Goudey <hans@blender.org>
Pull Request: https://projects.blender.org/blender/blender/pulls/147634
The final image produced by the compositor can have domain translation
on it (e.g. caused by a Translate or Transform node). Similar to how
the regular compositor viewer node remembers the output domain
translation, do the same in the compositor modifier.
Bubble back that translation up to VSE rendering code, where it is
then added to regular strip transform.
In order to make this "bubble up" part easier, refactored modifiers
so that instead of soup of parameters they all get a struct
ModifierApplyContext with all the relevant data.
Added a new VSE render test that covers various compositor
transformation nodes (translate, rotate, transform, corner pin).
Pull Request: https://projects.blender.org/blender/blender/pulls/147695
The fix in this case is to properly use the stored builtin attribute defaults
when capturing the field on the mesh. I extracted that to a function so
the code would read better with early returns.
Pull Request: https://projects.blender.org/blender/blender/pulls/147646
Support loops at the GLSL level instead of relying on
NOD_shader_nodes_inline.
This improves compilation and runtime performance, avoids causing
recompilations on iteration count changes, and allows supporting
dynamic iteration counts.
(EEVEE-only)
Pull Request: https://projects.blender.org/blender/blender/pulls/145269
Split the result mesh generation to work into two loops. The first loop
builds topology maps used for propagating attribute values, and the next
loop mixes attribute values with those maps. This has a few benefits:
The propagation can be done in parallel, and the topology mapping code
can get simpler. Also, we reduce the overhead that comes with iterating
over all attributes for every single element. Overall I didn't measure
a large performance difference though, because the new code has the
downside of needing to allocate two more arrays.
There is a small difference of a propagated byte color (250 vs. 251) in
the geometry nodes test due to a subtle difference in the mixing
implementation.
Part of #122398.
Pull Request: https://projects.blender.org/blender/blender/pulls/147367
The UI property was changed 7b97bc48d8
to a negative boolean but the boolean conversion inside EEVEE was
not inverted.
This mean that since 4.2, the default behavior for Lightprobe
volume has been broken / inverted.
To make an existing scene bake the same as before, all material
needs to have their `BackFace Culling > Light Probe Volume` options
inverted. This is done automatically through the versioning code.
The only test cases broken are the ones using default materials which
do not have their property turned off.
Release Notes should contains the compatibility breakage.
Pull Request: https://projects.blender.org/blender/blender/pulls/147218
This patch adjusts the LGG Color Balance mode to operate in linear space
as opposed to sRGB space, which is more inline with other software and
is more expected to the user.
Versioning was added using pre and post gamma correction, but it is
very slightly different in the blacks because the gamma node only does
power correction.
Pull Request: https://projects.blender.org/blender/blender/pulls/147533
Given a grid of velocities, this node advects the values of the input
grid. Different options for integration scheme and clamping are
exposed. Currently their names are just as defined in OpenVDB.
Pull Request: https://projects.blender.org/blender/blender/pulls/147273
Add a new constraint called "Geometry Attribute", which directly
samples vector, quaternion, or 4x4 matrix attributes from geometry and
applies these to an object's or bone's transform.
This can be used to transfer data generated by geometry nodes to the
object or bone level. By default the constraint will sample a vector
on the vertex domain. The default attribute is `position` for
simplicity, but the attribute value does not have to have anything to
do with neither the transform of the geometry object nor the geometry
itself.
Pull Request: https://projects.blender.org/blender/blender/pulls/136477
Currently when a strip has a transform that does not fill the whole
render area, first the image of the strip is transformed, and then
any modifiers are applied on that. This is mostly in the new
Compositor modifier, where procedural textures, gradients, image
coordinates "stick to the screen" instead of following the transformed
strip.
This changes the behavior so that first the modifiers are applied
to the strip image, and then the strip is transformed. This is
potentially a visually breaking change:
- This can alter visual look of existing strip, especially if they are
scaled. Previous behavior was first scale filtering, then modifier;
now it is first modifier, then scale filtering.
- Most obvious change is Compositor modifier (which is new in 5.0).
- Compositor modifier can actually expand the input image (e.g. Blur
node with "expand bounds" option set), and that works.
- Note that Masks continue to be applied in global/screen space. There
can be small look differences with rotated/scaled strips that use
masks, due to Mask application now needing to do filtered mask image
lookups.
- If anyone needs previous behavior (modifier is applied on the
"whole screen"), they can put transformed strip into a meta strip,
and apply the modifier on the meta strip itself.
Compositor modifier examples with images in the PR.
Pull Request: https://projects.blender.org/blender/blender/pulls/146181
This PR gives access to OpenVDB's level set filtering operations
through individual **SDF Grid** nodes.
- **Mean Curvature**: Smoothens high curvature areas more than flatter
ones.
- **Laplacian**: Approximates mean curvature flow for true SDFs at lower
computational cost.
- **Median**: Reduces noise while preserving sharp features.
- **Mean**: Fast separable averaging filter for general-purpose
smoothing with linear computational complexity.
- **Offset**: Uniform dilation/erosion operation that shifts the SDF
surface by a world-space distance.
- **Fillet**: Rounds off concave internal corners by operating only on
regions with negative principal curvature.
Pull Request: https://projects.blender.org/blender/blender/pulls/147224
The dilate threshold test fails on Qualcomm GPU due to an unknown
reason, but the tests shows significant differences in all platforms
regardless, they are just small enough to pass the test threshold. The
difference is most likely the new SMAA implementation done to match the
GPU version, so updating the reference is reasonable here.
Pull Request: https://projects.blender.org/blender/blender/pulls/147451
The result of the Keying Screen node tests is very different yet the
test succeeds, that's because the test result has a very small number of
pixels compared to the final image. To fix this, we match the size of
the result with the render size.
This also updated the test reference which should have been updated
after the new RBF interpolation implementation of the node.
Pull Request: https://projects.blender.org/blender/blender/pulls/147449
Previously, Repeat Zones with non-static iterations would be fully
disconnected.
With this patch, they behave as if they have 0 iterations.
Co-authored-by: Jacques Lucke
Pull Request: https://projects.blender.org/blender/blender/pulls/147211
Explicitly set the transform of a grid.
The new transform can fail to be applied if the input transform isn't
invertible or for some extremes of scaling (0 or combinations of
negative and positive) and numerical errors with `openvdb`. If a
transform is not applied an error is raised and the `Is Valid`
returns false.
Pull Request: https://projects.blender.org/blender/blender/pulls/146824
Adds four new grid operator nodes that wrap OpenVDB's differential
operators. All nodes take a grid input and output a grid, potentially
with a different data type.
New nodes:
- Grid Curl: Calculate curl of vector field (Vec3 → Vec3)
- Grid Divergence: Calculate divergence of vector field (Vec3 → Float)
- Grid Gradient: Calculate gradient of scalar field (Float → Vec3)
- Grid Laplacian: Calculate Laplacian of scalar field (Float → Float)
These operators enable vector calculus operations on volume grids for
effects like flow analysis, vortex detection, etc.
Co-authored-by: Hans Goudey <hans@blender.org>
Pull Request: https://projects.blender.org/blender/blender/pulls/146845
Adds a Join Bundle node with a multi-input Bundle socket. Bundles are
iterated on the top level of each input bundle and items added to the resulting
single bundle. While creating the final bundle existing items have priority and
new items with an already existing name are discarded.
There is an info message when there are duplicate keys in the input bundles.
Co-authored-by: Brady Johnston
Co-authored-by: Jacques Lucke
Pull Request: https://projects.blender.org/blender/blender/pulls/146750
When a strip modifier uses an adjustment layer that is above it
as the mask input, this leads to recursive rendering. Similar to the
fix in !146624, pass SeqRenderState to modifiers as well.
Pull Request: https://projects.blender.org/blender/blender/pulls/147029
This implements the Menu Switch node in shader nodes. It's the same node that is
used in Geometry Nodes and the Compositor.
The Menu Switch node is purely handled during preprocessing and thus builds on
top of #141936. Hence, it's input has to be a single value, just like the
iteration count for repeat zones. This limitation can be lifted in the future,
but currently there is no way to produce a non-single menu value in shader
nodes. This will become possible if other Switch nodes are added though.
Pull Request: https://projects.blender.org/blender/blender/pulls/146896
Applies thin film iridescence to metals in Metallic BSDF and Principled BSDF.
To get the complex IOR values for each spectral band from F82 Tint colors,
the code uses the parametrization from "Artist Friendly Metallic Fresnel",
where the g parameter is set to F82. This IOR is used to find the phase shift,
but reflectance is still calculated with the F82 Tint formula after adjusting
F0 for the film's IOR.
Co-authored-by: Lukas Stockner <lukas@lukasstockner.de>
Co-authored-by: Weizhen Huang <weizhen@blender.org>
Co-authored-by: RobertMoerland <rmoerlandrj@gmail.com>
Pull Request: https://projects.blender.org/blender/blender/pulls/141131
This node outputs tangent values for face corners. There are two methods:
- **Exact** is the same MikkTSpace calculation used elsewhere in Blender.
- **Fast** (from #131308) is over 4x faster, and useful in many of
the same situations, though not necessarily tangential to the surface.
The reason to include both methods is that there are use cases where the
quality of the tangents don't matter (though the results are actually very
similar visually), we just need some continuous values across faces.
Pull Request: https://projects.blender.org/blender/blender/pulls/145813
This PR enables strip curve drawing when performing the workbench
rendertests. On Intel/vulkan the lines are to far off. Using strip will
reduce platform differences. Downside is that (basic) line rendering is not
covered anymore by a render test.
Pull Request: https://projects.blender.org/blender/blender/pulls/146820
This adds support for packed linked data. This is a key part of an improved
asset workflow in Blender.
Packed IDs remain considered as linked data (i.e. they cannot be edited),
but they are stored in the current blendfile. This means that they:
* Are not lost in case the library data becomes unavailable.
* Are not changed in case the library data is updated.
These packed IDs are de-duplicated across blend-files, so e.g. if a shot
file and several of its dependencies all use the same util geometry node,
there will be a single copy of that geometry node in the shot file.
In case there are several versions of a same ID (e.g. linked at different
moments from a same library, which has been modified in-between), there
will be several packed IDs.
Name collisions are averted by storing these packed IDs into a new type of
'archive' libraries (and their namespaces). These libraries:
* Only contain packed IDs.
* Are owned and managed by their 'real' library data-block, called an
'archive parent'.
For more in-depth, technical design: #132167
UI/UX design: #140870
Co-authored-by: Bastien Montagne <bastien@blender.org>
Pull Request: https://projects.blender.org/blender/blender/pulls/133801
Previously it was only working for the single layer case. For multipart
we write the colorspace in each part. For single part we write the first
non-data colorspace, and hope data passes will be identified based on channel
name like Blender does (e.g. XYZ instead of RGB).
Reading is unchanged and still the same as before, in that it only reads the
colorspace from the first part. There is only one color space per image
datablock, so we can not store anything more currently. In practice it
would be unusual for all passes in a file not to either have the same
colorspace or be data.
All the compositor file output test images were updated to include the
metadata, so that the test will check if the metadata is there.
Ref #144911
Pull Request: https://projects.blender.org/blender/blender/pulls/146809
New "Interleave" option in image format settings, enabled by default in old
files and disabled by default in new files to write multi-part files.
By not storing all channels interleaved, it is possible for other applications
to load individual passes more efficiently. This might also work better with
compression.
We follow the OpenEXR docs in that the channel name is the full
Layer.Pass.Channel rather than just Channel. Some other renderers
(e.g. Houdini Karma) write just the channel and that is what our reading
code assumed so far.
I've verified that Nuke can read the multipart EXR files produced by Blender,
including multiview and cryptomatte.
Fix#123727
Pull Request: https://projects.blender.org/blender/blender/pulls/146650
* Add adaptive subdivision properties natively on the subdivision surface
modifier, so that other engines may reuse them in the future. This also
resolve issues where they would not get copied properly.
* Remove "Feature Set" option in the render properties, this was the last
experimental one.
* Add space choice between "Pixel" and "Object". The latter is new and can
be used for object space dicing that works with instances. Instead of
a pixel size an object space edge length is specified.
* Add object space subdivision test.
Ref #53901
Pull Request: https://projects.blender.org/blender/blender/pulls/146723
When effect of adjustment layer strip is moved below the adjustment
layer, this causes infinite loop in strip rendering. Same happens when
you use multicam strip and set source channel to one of its effects.
This is fixed by passing `SeqRenderState` to the effects. If any strip
renders "seqbase" pointer of strip is stored in set in the render state
struct. If the pointer exists in this set, function returns without
rendering anything. In other words, The strip must never render itself.
Pull Request: https://projects.blender.org/blender/blender/pulls/146624
This patch removes the XY scale inputs from the Displace node. The
inputs were redundant since they were just multiplied by the vector.
This simplifies the node and improves performance slightly.
Additionally, the Vector input was renamed to Displacement since it no
longer specifies a direction.
Pull Request: https://projects.blender.org/blender/blender/pulls/146356
This patch reorders the inputs of some of the compositor nodes
accordingly to their importance. The importance is already quantified
internally using the domain priority of the input, so we needn't make
any subjective judgement and just order by the priority.
This breaks forward and backward compatibility if input indices were
used as opposed to input identifiers due to the different order.
Handling compatibility is not impossible, but is difficult, and it was
already ignored in many past node changes.
Pull Request: https://projects.blender.org/blender/blender/pulls/146311
