Jump to content
  • Advertisement

Archived

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

HellRaiZer

Shadow Mapping Questions

This topic is 5415 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

A few days ago i started searching the net for shadow mapping tutorials and demos. I didn''t find much of them. If i remember correct there are 4. 3 from nVidia''s site, and 1 from somewhere else (i don''t remember where). First of all, let me say that i don''t have an ARB_shadow (and co) capable card. After some reading i found that the only way for shadow mapping with my card is "dual-texture shadow mapping". About the demos. The first of them was Everitt''s Projective Shadow Mapping demo from nVidia. I don''t like the quality of it''s shadows! They are blocky even with a 512x512 texture (yes i hacked the demo so it has 512x512 shadow maps). Also, the code isn''t the cleanest i''ve read (sorry Cass!!!). I abandonded it after a few hours. The second, also from nVidia, was the one which illustrates a few shadow mapping techniques (with or without extensions), but it was very complicated to follow. Finally the last demo from nVidia was the one i tried to figure out how it works, because the results were better than the other two, and because the code was written from a non-nVidia person!!! The author named the technique "Index Shadow Mapping". I modified it, so i can understand it, and after two days of trial and error, i finally made it work, without the extra things (reflections, refractions, etc.). I think i understand the basic structure of it (render the scene from light''s point of view, with depth writes and ONLY alpha channel enabled, store the result as the shadow map for that light, and when rendering the scene from the camera, project the shadow map onto geometry affected by the light, in a seperate pass using additive blending). One thing i noticed, and i don''t understand completely, is the alpha values (alpha index) he (the author) selects for the objects. If i have a bigger and more complex scene, how can i "generate" these values? An other strange thing is alpha bias. When i modified the original demo, playing around with the code and variables, alpha bias was standard (-0.5). When i transfered the rendering and initialization code to my framework, i noticed that this "standard" value wasn''t correct. I had to re-adjust it. The strange thing is that my app''s alpha bits are equal to demo''s alpha bits!!! Why might this happening? Is there an "equation" for calculating the exact value needed? Finally, i want to know if there is any tutorial/paper (not source without comments!) that describes the above technique with more details, or any tutorial in general for shadow mapping WITHOUT shadow extensions. Thanks in advance. HellRaiZer PS. The forth demo i found was about index shadow mapping with cubemaps (for point lights). I haven''t checked it yet, because i think it''s the same algorithm with the one i described above.

Share this post


Link to post
Share on other sites
Advertisement
Hello!
I don''t know if this helps you but check this site.
It''s about a new technique named Trapezoidal shadow maps and the results(from the downloadable videos) are pretty interesting. The paper about it hasn''t been unfortunately released yet. There''s also a link with explaination of their implementation of perspective SM.

Share this post


Link to post
Share on other sites
Thanks for the reply. I found what''s going on, with index shadow mapping on my own (after a big fight with the demo code).
If anyone is interested in the details i can post an overview.
I say this because i haven''t found any tutorial about this,so ...

There are still some unclear parts but i''ll manage to solve them.

Thanks again.

HellRaiZer

Share this post


Link to post
Share on other sites
Try http://www.paulsprojects.net/

There is an awsome tutorial on shadow mapping, and not index shadow mapping.

Share this post


Link to post
Share on other sites
Guest Anonymous Poster
i''ve recently developed a new simple approach for doing shadow mapping with increased resolution near the observer. i''ve described the algorithm on my homepage(code,math,pics) feel free to try it out
http://www.vmars.tuwien.ac.at/~scherzer/lssm/index.html

Share this post


Link to post
Share on other sites

  • Advertisement
×

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

We are the game development community.

Whether you are an indie, hobbyist, AAA developer, or just trying to learn, GameDev.net is the place for you to learn, share, and connect with the games industry. Learn more About Us or sign up!

Sign me up!