SSS distance all mapped to same channel + different scaling compared to Cycles

[gkn11]

The scattering distance for R, G and B seems to all be mapped to the x-channel of the radius vector instead of R - X, G - Y, B - Z. The factor also seems to be off by about 100 compared to cycles too.

Also, I'm not sure how the weight of the subsurface scattering works in the RPR Uber node. The documentation mentions that backsurface scattering is activated with SSS but it uses the SSS color instead. That results in if a weight of >0 && <1 then more light seems to pass through the object than compared to if the weight is 1.0. Not sure if this is a bug or intentional?

System Setup (Environment)

Arch Linux 6900XT, 5950X Blender 3.0.1

A small test scene is included with default Suzanne and a light behind to see how much light is visible through the ears. A radius vector with [0.02, 0.02, 0.02] for RPR, for Cycles you have to change the scaling to 100 to get a similar effect.

sss_suzanne.zip

[VascoPi]

Thanks need to be investigated.

[gkn11]

Did a few more tests and the scaling issue might be cycles using "random walk" as the default sss method where "scattering radius automatically adjusted to match color textures". Not sure how well that translates to RPRs method?