diffuse bug

[Alessandro Padovani]

daz studio 4.15, blender 2.91, commit c8826b0

I noticed that when there’s a diffuse roughness, the plugin plugs the diffuse map in the roughness channel, even if there’s no map in iray. This is true both for uber and pbrskin. Test scene attached, it’s a g8f with pbrskin, see the torso and face materials.

‌

‌

[Alessandro Padovani]

It also seems there’s a bug in bitbucket. When I edited #364 it created another issue #365, so I’m closing #364 since I can’t delete it.

[Alessandro Padovani]

Issue #364 was marked as a duplicate of this issue.

[Alessandro Padovani]

commit 62a8dc1 works fine

[Alessandro Padovani]

As for commit 2de1b94, it seems the diffuse channel is not imported when there’s dual lobe. This happens both with the uber and pbrskin shaders, both with the bsdf and principled options. Test scene included test2.duf, it’s a g8f with some pbrskin, see the face material for pbrskin and torso for uber.

‌

‌

‌

[Thomas Larsson]

This has never worked, I think. The Cycles and Eevee inputs of the node groups are linked to the previous node, and the outputs are linked to the next node. However, if Fac = 0, the node group is ignored, and if Fac = 1, the prevous node is not linked to it. This makes sense if the node group contains mix groups at the end, which mix the inputs with whatever is made inside the node group. However, the dual lobe group is more complicated, and you cannot ignore the inputs even if Fac = 1. So now the previous node is linked to the dual lobe.

It is a good idea to uncheck the Prune Nodetree global setting, to really see the nodes that are generated, before the last pass that removes the dead ends.

[Alessandro Padovani]

The fac pruning idea is good, but it can’t work for reflections, such as dual lobe and top coat, that are “added” features to a base surface, so the base surface has to be there. As for commit d4c1572 it works fine, marking as resolved.

Thank you Thomas for the fast fix.

[Thomas Larsson]

Fac pruning is implemented for top coat, but as far as I can tell it does work for top coat as it is now implemented. If it is replaced by something more complicated in the future, this must be reconsidered.