Add support for transmission_extra_roughness in Standard Surface

[niklasharrysson]

This change list adds support for parameter transmission_extra_roughness, that was previously unused in the MaterialX standard_surface implementation.

[niklasharrysson]

specular_roughness = 0.15, transmission_extra_roughness = 0.0

specular_roughness = 0.15, transmission_extra_roughness = -0.12 (removing most of the roughness from transmission)

specular_roughness = 0.15, transmission_extra_roughness = 0.05 (adding additional roughness to transmission)

[jstone-lucasfilm]

Looks good to me, thanks @niklasharrysson!

[dbsmythe]

I am thinking it might be better if this was implemented as a multiplier for the original roughness value, rather than being subtracted. In particular, if the original roughness value is texture mapped and contains a mix of very small and much larger values, then you couldn't subtract very much without roughness going negative. If implemented as a multipler this can never happen. I think that artistically, globally doubling or halving roughness on a texture-varying surface would look more natural as well.

[niklasharrysson]

Thanks for the feedback @dbsmythe.
I see your point, but the way it is implemented now, being additive/subtractive, is how it is defined in the Standard Surface whitepaper: https://autodesk.github.io/standard-surface/
It's how it has been implemented in Arnold and other Autodesk products. So if we want to change that behavior I think that would have to be discussed with Autodesk and the authors of the whitepaper.

[Minianikum]

I also see your point @dbsmythe, but I think that a multiplier would move away from the point of "transmission_extra_roughness".
The light passing through the material is scattered by the surface roughness, then it might also be additionally scattered by the medium itself - hence the "extra roughness" - that is not related to how rough the surface is, just added to it. If we were to use a multiplier the "blurriness" of the medium would always be proportional to the roughness of the surface, and there shouldn't be such a relation, they need to be separate things that are added together.
(they need to be capped to 0-1 of course)