Archived

This topic is now archived and is closed to further replies.

ktron

doom3 realtime lighting?

Recommended Posts

hey, i''m just kinda wondering about how carmack is doing that amazing lighting in doom3? Is he using something comparable is direct3d''s shadow volume''s with everything in the game or is he doing something completly different? Just Wondering, Kevin

Share this post


Link to post
Share on other sites
No, I don''t think it''s truform tech since truform is ati''s creation. Also, I think I read in an interview with John Carmach that doom3 wouldn''t be complatable with truform. (?)

Anyone else?

thanks,
Kevin

Share this post


Link to post
Share on other sites
quote:
Is he using something comparable is direct3d''s shadow volume''s...

Shadow volumes (note the correct pluralization) are not intrinsic to Direct3D or any other 3D API. I hope you''re aware of that.

Share this post


Link to post
Share on other sites
Merlin, dude. What''re you? Department of Corrections? I don''t usually respond to say nothing, but why didn''t you correct the whole sentence while you were at it (should have used a ''to'' instead of ''is'')?

Or...you could''ve just told him that they''re not exclusive to DirectX and left it at that. Providing new information to the uninformed is laudable, but making corrections just to show how smart you are is laughable. I hope you''re aware of that.

Share this post


Link to post
Share on other sites
I don't follow what John Carmack does very closely but I'm sure you can find plenty of info on what he's doing for Doom 3 in some interviews he did post-E3. You might start here: http://www.beyond3d.com/interviews/carmackdoom3.

From what I've seen it looks like per-pixel lighting of the type that MS and NVidia have been selling to ISVs for years now. I'm sure you can find papers on NVidia's site explaining how to do this using D3D and OpenGL.

Truform is a tesselation technology that belongs to ATI. The reason John claimed he didn't want to use it was because his engine is calculating shadow volumes dynamically and all his engine has access to is the low-poly count (pre-tesselation) models. Therefore, even though the model would look great on screen the shadow volume would not match the model.



[edited by - don on August 11, 2002 11:31:58 PM]

Share this post


Link to post
Share on other sites
Ok. Cool, although I was kinda wondering how exaclly he did it, which hasn''t really been talked about in the interviews I''ve seen. does he render the scene from the view of each of the lights in a scene and use that data to fill up the camera view with shadow data?

Share this post


Link to post
Share on other sites
Im pretty sure he does standard shadow-volume building, by rendering the scene from each lightsources point of view, yes.


--
MFC is sorta like the swedish police... It''''s full of crap, and nothing can communicate with anything else.

Share this post


Link to post
Share on other sites