On 03/11/2011 10:12 AM, Stephen wrote:
> What I think we need to do is scale both the subsurface values and the
> texture, independently but related. (If that makes sense)

agreed ... I'm doing a beauty run on the stanford dragon now, here's
what I have for the material:

object { stanford_dragon
material {
  texture {
    T_Stone18
    scale 0.035
  finish {
    subsurface {translucency <0.05,1,1>*0.5}
    reflection { SeaGreen*0.001 }
    }
  }
  interior {ior 1.62}
  }

I copied T_Stone18 into my scene file and made a couple of mods (ambient
to 0 and dialed down diffuse a touch) in the bottom layer texture, then
changed phong to specular/roughness on the outer layer.

To setup the sslt I used a plain white texture instead of T_Stone18
while I dialed in the sslt (seems to help with visualizing the effect).
That was interesting because my 1st take on mm_per_unit was that it
would be smaller than default because of the size of the model at scale
1. When smaller didn't seem to be working I tried default without much
luck either, so I increased to 1000 then 2540, but still had to wildly
scale the translucency color. Finally I decided to try a couple of half
step itterations between something I considered to be more reasonable
(100 - 10) and found the sweet spot. The value I arrived at was so close
to the subsurface scene file values, I opted to use mm_per_unit 40 and
quickly was able to tune the translucency setting to what's shown above.

> PS I notice that you forgot to put your subsurface values in the RC3
> format :-P

ha-ha ... bleeding edge :-P

Post a reply to this message