Using BRDF measurement representations with Radiance

radiance-general
1

Hi

I'm currently working on how to use measured BRDF data, e.g. represented using a wavelet parameterization, with Radiance. I'm struggling at bit with this, as documentation on using Radiance with measured BRDF data seems not to exist? Can someone give me directions, links to previous work on this, documentation or advice on this?

Thanks,
Jeppe Vesterbaek
Univesity College London

2

Hi Jeppe,

Check out "He.cal" in the ray/lib directory of the standard distribution. It's a bit complicated, but you'll probably end up with something similar if you're building a parametric BRDF model.

You should be aware that Radiance doesn't do all it should with BRDF materials, in that the indirect component is approximated as diffuse. Only highlights use the BRDF, not secondary rays, for efficiency.

-Greg

···

From: Jeppe Vesterbaek <[email protected]>
Date: October 5, 2005 10:21:19 AM PDT

Hi

I'm currently working on how to use measured BRDF data, e.g. represented using a wavelet parameterization, with Radiance. I'm struggling at bit with this, as documentation on using Radiance with measured BRDF data seems not to exist? Can someone give me directions, links to previous work on this, documentation or advice on this?

Thanks,
Jeppe Vesterbaek
Univesity College London

3

Sorry, but this is one of those design decisions that is built into the code. Of course, the C source is all there if you want to modify it, but there's quite a lot to figure out along the way, like how to sample the hemisphere for an arbitrary BRDF. This is a difficult problem, and one that has been tackled by various people over the years, so there are some precedents.

If I were to do it myself, I would investigate something like a median cut algorithm or kd-tree on the hemisphere, which would require presampling the BRDF or performing lazy evaluation. For something like the He-Torrance model, you would likely spend hours and hours (perhaps days) just computing the BRDF before you ended up tracing your first ray. This is why Radiance doesn't make the attempt.

-Greg

···

From: Jeppe Vesterbaek <[email protected]>
Date: October 5, 2005 11:12:03 AM PDT

Hi Greg

Thanks for the quick reply. I'll have a look at "He.cal" tomorrow (been staring too long at this monitor by now :p).

With respect to Radiance's handling of BRDF materials: is there no way (around) to change how Radiance calculates the indirect component? I'll be using Radiance for rendering (relighted) real scenes (in collaboration with Katrien Jacobs) and would like the rendering to be as realistic as possible (no efficiency concerns atm.).

Thanks,
/Jeppe