I have difficulties to understand the roles of the different transmission-related material properties I get when distilling to transmissive_pbr. I’m currently implementing transparency rendering in a pbr renderer for realtime rasterization. I can imagine the purpose of transparency (in the formula below referenced as t), transmission_color (tc), attenuation_distance (ad) and attenuation_color (ac), but I am not sure. What I can imagine is the following, where s is the surface color returned by the lighting calculations respecting roughness, normals etc, b the background color behind the object/volume and d the depth or thickness of the object/volume in view direction:
finalColor = s*(1-t) + t * tc * (ac * ad/d + b*(1- ad/d ))
Is this correct?
But I don’t know, in which relation subsurface_color, anisotropy and anisotropy_rotation should be integrated. How could I do that?
Right now I don’t want to use the IOR.
Hints to helpful literature are as appreciated as suggestions how to solve this.
Anisotropy and anisotropy rotation are not aspects of the transmission here but aspects ofthe roughness of the glossy bsdf. Sorry for the late answer, i will look and verify the formula
Thanks for the answer. My formula above is definitely wrong for different reasons. For example I accidentally wrote ad/d instead of d/ad and it returns the attenuation color even if there is no light coming from anywhere. That way it is giving the transparent objects a ghost-like volumetric look, which looks pretty but is completely wrong.
What I have now is the following:
finalColor = (tc * b * (1-d/max(ad,d)) + b * ac * d/max(ad,d)) * t + s * (1-t)
It looks better, but I’m still unsure about the correctness and the subsurface color is still not included. Right now the formula is also ignoring reflections inside the objects, because I’m still working on an SSR-similar way to compute the length of the reflected ray inside the object.
for subsurface color, you get the absorption coefficient in the material. if you have the length of travel for the light through the medium, then the attenuation factor would be
attenuation = exp(-absorption_coefficient*distance_traveled)
Hi,
I recently found something very similar, good to see this confirmed. Thanks a lot!