Closed UX3D-haertl closed 4 months ago
This is an IBL "furnace test" where the same amount of light comes from all directions. It's hard to read but ideally all of the same-color spheres would be the same brightness here, as they share the same base color.
On the left is main
, on the right is this PR, using MetalRoughSpheres with this IBL:
EdM_75percent_furnace.zip
The darkness in the upper-right shows some loss of energy with highly rough metallic spheres. This was already a known problem and has got a bit worse on this branch I think.
Also some darkening is present along the middle horizontal row, between dielectric and metallic. I wonder if this is an unintended consequence of recent changes to the sample implementation.
Implements the fixes discussed here: https://github.com/KhronosGroup/glTF/issues/2386 Fixes #530
Known issues which were already there before the refactoring:
552
Comparison of previous state and new one. Some assets are the same, but most assets got darker with IBL and brighter with punctual lights: brdf_refactor.zip