Rough Specular reduces specularity amount?

[adanmq]

Are there photos of the actual material that was scanned? I have no idea which one of these is closest to the intended result.

Best regards,
Vlado

There is a preview render inside the pack but since i don´t know the original light conditions i can´t be 100% sure this preview match reality. My test it´s just to see fresnel differences across shaders.



Best.

[vlado]

Thanks! It would be nice if they provide some kind of more precise reference for the materials (like a sample scene or something); it seems like it's needed, what with the different renderers interpreting things in different ways (and even the same renderer with different materials ) Disney never provided a reference implementation of their principled BRDF, I think, and so many parts of it are open to interpretation...

Best regards,
Vlado

[^Lele^]

Their Paper (pbs_2012) is sketchy on the implementation details, although if one reads accurately through there are some hints scattered here and there.
The Rman implementation is possibly the closest one to the original paper, i'd think, as it's all in house, if in different offices.

This said, taking the Disney shader as PBS yardstick is risible, imho.
I have hardly ever seen such a hack-job of a shader, with no energy conservation, arbitrarily clamped values for parameters and no refraction support whatsoever.

Debate, if you know what i'm talking about.

[FSGFX]

the megascan samples are made in realtime with there tool... no offline render...

[vlado]

the megascan samples are made in realtime with there tool... no offline render...

Do they have that tool available anywhere?

Best regards,
Vlado

[FSGFX]

yes if you have an subscription...
https://megascans.se/faq#What%20is%2...ns%20STUDIO%3F

[^Lele^]

https://www.youtube.com/watch?v=sNLOBO1hrpo

https://www.youtube.com/watch?v=B2fT654ajZg

The two tools they provide (for a sum.).

[Richard7666]

Yep!

In current shading models they consider when doing rough / microfacet surfaces they have an equation that just cuts out a portion of the light that gets trapped in the shadows of the rough surface whereas a lot of it would end up being directed into the surface itself and doing sss things! For something like chalk there's probably even a tiny amount of GI going on with light bouncing off the surface and then pinballing between some of the microfacets, giving it an almost glowy effect!

Here is an example of what I am thinking of. It seems visible in the sides that are indirectly illuminated

[^Lele^]

the exact term is retro-reflection, i believe.
you can achieve that easily with GTR, by lowering the gamma and gloss (without fresnel, nor the need for a diffuse.).

[Richard7666]

the exact term is retro-reflection, i believe.
you can achieve that easily with GTR, by lowering the gamma and gloss (without fresnel, nor the need for a diffuse.).

to clarify, do you mean basically:

-turn off fresnel reflections/any falloff map acting as fresnel etc
-set refract glossiness to something very low so that it 'spreads' over the whole material
-but then I'm not sure where the gamma part comes into it?? Sorry!

[^Lele^]

-but then I'm not sure where the gamma part comes into it?? Sorry!

Correct for the first two.
Sorry for my misuse of terms.
I meant GTR lobe to tail ratio (just under the BRDF selection dropdown).

[Midiaeffects]

Yep!

In current shading models they consider when doing rough / microfacet surfaces they have an equation that just cuts out a portion of the light that gets trapped in the shadows of the rough surface whereas a lot of it would end up being directed into the surface itself and doing sss things! For something like chalk there's probably even a tiny amount of GI going on with light bouncing off the surface and then pinballing between some of the microfacets, giving it an almost glowy effect!

AFAIK, the Multiscatter with Smith Model is supposed to solve this energy loss.

https://eheitzresearch.wordpress.com/240-2/

-Eugenio

[sharktacos]

Disney never provided a reference implementation of their principled BRDF, I think, and so many parts of it are open to interpretation...

They provide a GLSL version of the shader. Would that work?

[vlado]

They provide a GLSL version of the shader. Would that work?

Hm, where is it?

Best regards,
Vlado

[FSGFX]

someone on polycount uploaded the .3do scenes for megascans... (you need ddo to open them)
an automated process is creating the shaderballs with those files...

i will send a mail in your direction...