Computer Graphics and Haskell - Radiosity Methods

Hello cafe, While I was studying for my computer graphics test I have tomorrow I realized that maybe some of the major problems I've read so far about Radiosity Rendering Algorithms may be reduced significantly if it was implemented in Haskell and taking advantage of the lazy evaluation so that only what can be seen from the viewer's perspective point of view is calculated, and the rest of the scene just remains as thunks waiting for them to be calculated in case they are needed. I guess that probably most of the people on this list doesn't really work on the computer graphics area, but if anybody out there actually does I'd really like if he or she could explain me more of why what I'm saying could be really useful and what has already been done (or why I'm so lost that I'm probably gonna fail my exam tomorrow hahaha)... One of the things I'm studying is the Monte Carlo Radiosity Methods which takes advantages of non-determinism to create stochastic algorithms, I guess this is more difficult to implement in Haskell, please if you know any of this Monte Carlo Methods for Radiosity Algorithms share that knowledge with me (and the list). Thank you, Hector Guilarte

you could have a look at these ... bjpop-ray - search on web hray - http://users.elis.ugent.be/~kehoste/Haskell/HRay/
Hello cafe,
While I was studying for my computer graphics test I have tomorrow I realized that maybe some of the major problems I've read so far about Radiosity Rendering Algorithms may be reduced significantly if it was implemented in Haskell and taking advantage of the lazy evaluation so that only what can be seen from the viewer's perspective point of view is calculated, and the rest of the scene just remains as thunks waiting for them to be calculated in case they are needed.
I guess that probably most of the people on this list doesn't really work on the computer graphics area, but if anybody out there actually does I'd really like if he or she could explain me more of why what I'm saying could be really useful and what has already been done (or why I'm so lost that I'm probably gonna fail my exam tomorrow hahaha)...
One of the things I'm studying is the Monte Carlo Radiosity Methods which takes advantages of non-determinism to create stochastic algorithms, I guess this is more difficult to implement in Haskell, please if you know any of this Monte Carlo Methods for Radiosity Algorithms share that knowledge with me (and the list).
Thank you,
Hector Guilarte
------------------------------------------------------------------------
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

2010/3/1 Hector Guilarte
Hello cafe, While I was studying for my computer graphics test I have tomorrow I realized that maybe some of the major problems I've read so far about Radiosity Rendering Algorithms may be reduced significantly if it was implemented in Haskell and taking advantage of the lazy evaluation so that only what can be seen from the viewer's perspective point of view is calculated, and the rest of the scene just remains as thunks waiting for them to be calculated in case they are needed.
I don't think that you really can get away without calculating those out-of-view parts of scene. But you certainly can make rendering slightly easier (or not) by using infinite scene subdivision trees and walk them lazily.

2010/3/1 Hector Guilarte
: Hello cafe, While I was studying for my computer graphics test I have tomorrow I realized that maybe some of the major problems I've read so far about Radiosity Rendering Algorithms may be reduced significantly if it was implemented in Haskell and taking advantage of the lazy evaluation so that only what can be seen from the viewer's perspective point of view is calculated, and the rest of the scene just remains as thunks waiting for them to be calculated in case they are needed.
The way radiosity works, those invisible parts of the scene can still add illumination to the visible parts. So the first time you query the radiosity data for any part of the scene, you'll end up forcing the calculation of the entire radiosity solution. That's basically the difference between plain raytracing, which is lazy in that way and works backwards from the viewer's perspective, and radiosity. Maarten
participants (4)
-
Hector Guilarte
-
John Lask
-
Maarten Hazewinkel
-
Serguey Zefirov