Just a thought, but something radiosity-like may be possible with the
existing photon patch (It should be noted though that I do not have the
photon patch because of a lack of processing power, I don't even have
the patience to do extensive traditional ray tracing on my PC. However,
I have followed the developement of the patch and have a pretty good
idea of how it works).

Here goes:
The diffuse interreflection simulated by radiosity and Pov Ray's
distributed raytracing is caused by imperfections on the reflecting
surface scattering the light in every direction. So why not do just
that. By creating a surface normal of something like bumps and scaling
the pattern to a very small scale, it may be possible to have the
reflected photons scatter to illuminate the scene. This has the one
disadvantage of requiring a large number of photons. HOWEVER, as I
understand it, the photon patch uses a variable to control how many
photons are gathered from the neighbouring regions to determine the
illumination of a point. Since global illumination does not require the
same precision as caustics, it may be possible to increase the area that
photons are gathered from and therefore decrease the total number of
photons required for the image to look globally illuminated.

Wait! It's not as crazy as it sounds!
I once used a similar technique to fake blurry reflections and the
results were pretty good. Also, this is essentially a modification of
what Pov Ray already uses to do global illumination.

Well, that's just a thought, anyone care to give it a shot?

BTW: Before the photon patch, people were always complaining that they
couldn't do disco balls. Now we have the means but I have yet to see one
in p.b.i, what gives?

-- 
Chris Maryan
mailto:cma### [at] geocitiescom
***
Will work for cash.
***
Any ideas presented in this message
are mine, mine and all mine. So if you
find a way of profiting from them, I
demand recognition and a cut of the pie.
Gimme, gimme, gimme.

Post a reply to this message