[Bf-committers] Realistic materials question (Farsthary)

[Yves Poissant]

From: <echelon at infomail.upr.edu.cu>
Sent: Monday, February 23, 2009 3:18 PM

>  Modified Blinn-Phong model--->  kd/PI + ks*(n+2)*(nor,H)^n/(2*PI) also
> have 3 parameters

This modified model already includes the normalizing of the specular 
component. Now, all you need, is make sure ks+ks <= 1.0 the the equivalent 
raytracing render. The path tracer will take care of that by itself with the 
Russian roulette.

>  what I mean is what are the ever needed material properties for a BxDF
> model? here I'm very open to ideas, suggestions,papers,links and so :)

For a BRDF, the ever needed material property would be roughness (two 
roughnesses for anisotropy). From roughness (which is roughly the reciprocal 
of the specular exponent), you would derive kd and ks for your russian 
roulette. One simple way of deriving that would be Schlick "An Inexpensive 
BRDF Model for Physically-Based Rendering".

>  also I need good formulas for implementing tone mappers because path
> tracer images fall outside the (0,1) range.

While it is nice to have a tone mapping algo, you can run quite some mileage 
with a very simple multiplier (for controling the exposure) followed by a 
simple gamma correction. This simple scheme has the added advantage that you 
don't have to wonder at the (sometime quite weird) idiosyncrasies of the 
tone mapping algo. Whatever you choose, I would suggest you save your 
renders to a floating point file format such as OpenEXR. Then, you can 
experiment with different tone mapping algos. For the algos, here is a good 
place to start:
http://www.mpi-inf.mpg.de/resources/tmo/

Yves