Cycles oneAPI kernel library was compiled using -ffast-math. The current
version of Clang makes it link to crtfastmath.o in that case, bringing a
static constructor that does set the FTZ/DAZ bits in MXCSR for the whole
program, leading to unwanted behavior with other components.
Instead of -ffast-math, we switch to a safer subset of compile flags.
Pull Request: https://projects.blender.org/blender/blender/pulls/111708
based on concentric disk mapping.
Concentric disk mapping was already present, but not used everywhere.
Now `sample_cos_hemisphere()`, `sample_uniform_hemisphere()`, and
`sample_uniform_cone()` use concentric disk mapping.
This changes the noise in many test images.
Pull Request: https://projects.blender.org/blender/blender/pulls/109774
This reverts commit 206ab6437b.
Seems that the illegal address error should be covered elsewhere, but it's not directly
clear where. Revert the commit for further investigation.
Discovered during an investigation into #111277
in rare situations (E.G. When normals are NaN), an emitter
won't be selected as part of `light_tree_cluster_select_emitter()`
and as a result of that, an `emitter_index` of `-1` is passed to
`kernel_data_fetch(light_tree_emitters, emitter_index)` resulting in
an "illegal address" error on some devices.
Pull Request: https://projects.blender.org/blender/blender/pulls/111292
This is in prevision of EEVEE panoramic projection support.
EEVEE-Next is planned to add support for these parameters.
Not having these parameters in Blender DNA will make Cycles
and EEVEE not share the same parameters and will be confusing
for the user.
We handle forward compatibility by still writing the parameters
as ID properties as previous cycles versions expect.
Since this change will break the API compatibility it is crucial
to make it for the 4.0 release.
Related Task #109639
Pull Request: https://projects.blender.org/blender/blender/pulls/111310
Fixes NaN in Vector Displacement node caused by the normalization of
0, 0, 0 vectors.
This fixes both visual rendering issues and an "illegal address" error
on the GPU. The "illegal address" error came from the Light Tree
Sampling code not handling the NaN normals well, leading to weird code
paths being taken, eventually leading to a kernel_assert and a
user facing illegal address error.
Pull Request: https://projects.blender.org/blender/blender/pulls/111294
Implements the paper [A Microfacet-based Hair Scattering
Model](https://onlinelibrary.wiley.com/doi/full/10.1111/cgf.14588) by
Weizhen Huang, Matthias B. Hullin and Johannes Hanika.
### Features:
- This is a far-field model, as opposed to the previous near-field
Principled Hair BSDF model. The hair is expected to be less noisy, but
lower roughness values takes longer to render due to numerical
integration along the hair width. The hair also appears to be flat when
viewed up-close.
- The longitudinal width of the scattering lobe differs along the
azimuth, providing a higher contrast compared to the evenly spread
scattering in the near-field Principled Hair BSDF model. For a more
detailed comparison, please refer to the original paper.
- Supports elliptical cross-sections, adding more realism as human hairs
are usually elliptical. The orientation of the cross-section is aligned
with the curve normal, which can be adjusted using geometry nodes.
Default is minimal twist. During sampling, light rays that hit outside
the hair width will continue propogating as if the material is
transparent.
- There is non-physical modulation factors for the first three
lobes (Reflection, Transmission, Secondary Reflection).
### Missing:
- A good default for cross-section orientation. There was an
attempt (9039f76928) to default the orientation to align with the curve
normal in the mathematical sense, but the stability (when animated) is
unclear and it would be a hassle to generalise to all curve types. After
the model is in main, we could experiment with the geometry nodes team
to see what works the best as a default.
Co-authored-by: Lukas Stockner <lukas.stockner@freenet.de>
Pull Request: https://projects.blender.org/blender/blender/pulls/105600
The `Find*.cmake` modules originally used uppercase commands to match
CMake's own conventions. Since then CMake uses lower-case and even
within our own find modules, using all uppercase wasn't done
consistently. Opt for lowercase everywhere.
This PR adds the Lacunarity and Normalize inputs to the Noise node
similar to the Voronoi node.
The Lacunarity input controls the scale factor by which each
successive Perlin noise octave is scaled. Which was previously hard
coded to a factor of 2.
The Noise node normalizes its output to the [0, 1] range by default.
The Normalize option makes it possible for the user to disable that.
To keep the behavior consistent with past versions it is enabled by
default.
To make the aforementioned normalization control easer to implement,
the fractal noise code now accumulates signed noise and remaps the
final sum, as opposed to accumulating positive [0, 1] noise.
Pull Request: https://projects.blender.org/blender/blender/pulls/110839
In the commonly used cycles headers, it's enough to include
much smaller <iosfwd> than the full <iostream>. While looking at it,
removed inclusion of some other headers from commonly used headers,
that seemed to not be needed.
Pull Request: https://projects.blender.org/blender/blender/pulls/111063
Overall, this commit reworks the component layering in the Principled BSDF
in order to ensure that energy is preserved and conserved.
This includes:
- Implementing support for the OSL `layer()` function
- Implementing albedo estimation for some of the closures for layering purposes
- The specular layer that the Principled BSDF uses has a proper tabulated
albedo lookup, the others are still approximations
- Removing the custom "Principled Diffuse" and replacing it with the classic
lambertian Diffuse, since the layering logic takes care of energy now
- Making the merallic component independent of the IOR
Note that this changes the look of the Principled BSDF noticeably in some
cases, but that's needed, since the cases where it looks different are the
ones that strongly violate energy conservation (mostly grazing reflections
with strong Specular).
Pull Request: https://projects.blender.org/blender/blender/pulls/110864
Many calls to add_check_c_compiler_flag add_check_cxx_compiler_flag
resulted in over long lines & visual noise. Replace with a function that
takes multiple (cache_var flag) pairs to reduce duplication.
Both the `Math` node and the `Vector Math` currently only explicitly
support modulo using truncated division which is oftentimes not the
type of modulo desired as it behaves differently for negative numbers
and positive numbers.
Floored Modulo can be created by either using the `Wrap` operation or
a combination of multiple `Math` nodes. However both methods obfuscate
the actual intend of the artist and the math operation that is actually
used.
This patch adds modulo using floored division to the scalar `Math` node,
explicitly stating the intended math operation and renames the already
existing `"Modulo"` operation to `"Truncated Modulo"` to avoid confusion.
Only the ui name is changed, so this should not break compatibility.
Pull Request: https://projects.blender.org/blender/blender/pulls/110728
The current core was only detecting sRGB transform when it is defined
as sRGB->Linear using srgb.spi1d. If it is defined as an inverse of
Linear->sRGB using srgb_inv.spi1d, or as an analytical formula using
ExponentWithLinearTransform then the code did not detect the color
as sRGB on anything by Apple Silicon platform.
The naming of the checks could be improved to make it more clear that
the check is only used to allow lossless access to 8bit sRGB textures.
Ref #110685
Pull Request: https://projects.blender.org/blender/blender/pulls/110889
This pull request covers up a subtle difference between the CPU and GPU
when rendering with a light tree. Specifically a case where the user
has a sun light with a small angle.
The difference was caused by the dot() function being different between
CPU and GPU backends, with the GPU showing more meaningful
floating-point precision losses when working with small suns.
Pull Request: https://projects.blender.org/blender/blender/pulls/110307
The issue was an out-of-bounds read access when checking whether
the world volume emission needs to be accumulated.
Solution is to check for this case. Done in the generic place, so
that the shade_volume kernel is more readable and no branching
added there, and there is no impact on scenes without the light
linking.
Assume that the world emissive volume belongs to the default light
linking group, as there is no way to link it explicitly to anything.
Pull Request: https://projects.blender.org/blender/blender/pulls/110733
This replaces the Sheen model used in the Principled BSDF with the
model from #108869 that is already used in the Sheen BSDF now.
The three notable differences are:
- At full intensity (Sheen = 1.0), the new model is significantly
stronger than the old one. For existing files, the intensity is
adjusted to keep the overall look similar.
- The Sheen Tint input is now a color input, instead of the
previous blend factor between white and the base color.
- There is now a Sheen roughness control, which can be used to
tweak the look between velvet-like and dust-like.
Pull Request: https://projects.blender.org/blender/blender/pulls/109949
