View previous topic :: View next topic |
Author |
Message |
mrshoe
Joined: 12 Aug 2005 Posts: 6 Location: Palo Alto, CA
|
Posted: Fri Aug 12, 2005 3:01 pm Post subject: Sol's Goals |
|
|
Sol aims to be a sophisticated global illumination raytracer. Some of the main goals of the project are:
* Raytrace very complex scenes in a reasonable amount of time.
* Simulate all light in a scene.
* Experiment with light transport in complex materials.
* Maintain a clear and elegant architecture and coding style.
If you are interested in computer graphics and would like to get involved in Sol's development, let me know. |
|
Back to top |
|
|
h3r3tic
Joined: 30 Mar 2004 Posts: 261 Location: Torun, Poland
|
Posted: Thu Aug 25, 2005 9:10 am Post subject: |
|
|
Cool project I'm looking forward to seeing it developed more. I'm myself somehow into global illumination, but my current approach is precalculating the indirect diffuse illumination via photon mapping and storing it in mesh vertices and rendering via OpenGL. Some (old) screenshots are here:
http://codeinsane.info/listDir.spy?viewType=img&path=code?2Fheresy?2Fimages
/+ that cornell box was calculated using a finite element radiosity method, not photon mapping. I've got much nicer screenies which I'll upload later +/
greetz ! </h3> |
|
Back to top |
|
|
mrshoe
Joined: 12 Aug 2005 Posts: 6 Location: Palo Alto, CA
|
Posted: Thu Aug 25, 2005 2:18 pm Post subject: |
|
|
Very cool!
I'm glad to see that there is interest in graphics in the D community.
I plan on implementing path tracing and photon mapping to accomplish global illumination in Sol, with the latter possibly accelerated by irradiance caching.
With speed not being the primary goal, I think I'm going to stick to ray tracing for the rendering to simplify things.
Of course the current effort is to build an efficient BSP tree! (Hopefully this will be checked in soon)
Your input is welcome. |
|
Back to top |
|
|
h3r3tic
Joined: 30 Mar 2004 Posts: 261 Location: Torun, Poland
|
Posted: Mon Aug 29, 2005 6:47 am Post subject: |
|
|
I suggest a kD-Tree.. and reading Jensen's book about photon mapping (or parts of it). I'll upload my new sources shortly to dsource so you can check my photon map, kdtree, bsp and accurate polygon splitting implementations. I think you could borrow a lot of code from me |
|
Back to top |
|
|
h3r3tic
Joined: 30 Mar 2004 Posts: 261 Location: Torun, Poland
|
|
Back to top |
|
|
mrshoe
Joined: 12 Aug 2005 Posts: 6 Location: Palo Alto, CA
|
Posted: Mon Aug 29, 2005 2:38 pm Post subject: |
|
|
Thank you, h3r3tic.
I read Henrik's book when I was taking his course at UCSD -- a good read indeed. I implemented most of what he describes in the book, but I was able to use his C++ implementation of photon mapping, including his kd-tree. I'll definitely check out your code before I implement it in D.
He also helped me implement basic subsurface scattering based on the photon map. I really want to tackle his fast BSSRDF method in Sol, but that will require a deeper understanding of the math behind it than I currently possess, I'm afraid. |
|
Back to top |
|
|
h3r3tic
Joined: 30 Mar 2004 Posts: 261 Location: Torun, Poland
|
Posted: Mon Aug 29, 2005 3:24 pm Post subject: |
|
|
Damn, I'd love to attend one of his courses But being located in Poland, that's rather hard I might actually use some of your help if I find myself troubled with something within global illumination In the meantime you can check out my new engine demo (look at my forum at dsource ). |
|
Back to top |
|
|
|