@david494 I believe your suggestion would fix #114940, would you like to make a PR and change BSDF_ROUGHNESS_SQ_THRESH to 2e-10?
Metal assumes the normal is already normalized:
METAL_FUNC float3 reflect(float3 i, float3 n)
{
return i - float(2) * dot(n, i) * n;
}
What's the motivation of this fix? Would…
For the bug fixes we put the Fix # in the commit title
I usually add # in the commit title, but this commit title is too long for that and I can not think of a good…
Does clamping cos_NH2 to 1 or returning M_1_PI_F / alpha2 when cos_NH2 is close to 1 fixes the issue?
I'm not in favor of increasing the threshold due to #114940.
The problem is demonstrated by the attached transparent_hits.blend. On CPU INTEGRATOR_SHADOW_ISECT_SIZE is 1024, larger than the maximal transparent bounces in the scene, so shadow_all() has…
This appears to be the same problem as #125375, and I can confirm applying 8a812e334d fixes the issue. It's already on the [backport list](blender/blender#12445…
The second commit is a refactor and can be reviewly separately. I will not squash the commits.
Each location in an overlapping region is composed of x_1\% of volume 1 and x_2\% of volume 2, so the phase function at that point should be x_1\%p_1+x_2\%p_2 instead of (p_1 +…