Open emackey opened 2 years ago
Earlier conversation about this topic can be found here: https://github.com/KhronosGroup/glTF-Sample-Models/pull/319#issuecomment-876812842
For reference, I tried to display this model in some libraries.
I thought the change from Smooth
toRough
in three.js was a bit steep.
TransmissionRoughnessTest.gltf result: |
---|
![]() |
![]() |
![]() |
![]() |
![]() |
Edit 2021.08.20 I added the display result in Dassault path-tracer.
@emackey : Is there anything to do on this issue? It is not clear to me.
If there are no responses to this comment, the issue will be closed by 27 Nov.
@DrX3D I suppose this issue could be transferred to the new repo. Sadly this isn't high-priority given my current project list at the moment.
It is not clear to me.
This is an epic pun on an issue about a transmissive model with too much roughness.
It was noticed that path tracers and rasterizers see different amounts of roughness on a clear volume, given that the path tracer will take into account roughness on the back of a transmissive surface, but a rasterizer will not.
We could consider adjusting the TransmissionRoughnessTest model to remove the roughness from the backs of the sample tiles, so all the roughness is on the fronts only. This might allow a more direct comparison between the different rendering strategies.