• Advertisement
Sign in to follow this  

Question about ID3DXPRTEngine::ClosestRayIntersects

This topic is 3383 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

Hi, I'm working on a PRT project, but unfortunately I'll have to wrote my own direct/indirect simulation code instead of using ID3DXPRTEngine::ComputeDirectLightingSH and ID3DXPRTEngine::ComputeBounce directly due to the lack of source code on these functions. So I reimplemented the 2 functions. There are 2 questions. After I tried my own code with the same number of Monte Carlo rays, there are more visible artifacts than what I get with that of the SDK's, while my code runs faster. I think the SDK's functions may have used some importance sampling scheme in order to avoid under sampling? Particularly I used 1024 rays. I used the functions ID3DXPRTEngine::ShadowRayIntersects and ID3DXPRTEngine::ClosestRayIntersects for ray tracing, which work reasonably fast as I expect since the DX PRT engine probably has used spatial subdivision for ray tracing accelerations. However, I got severe aliasing along the mesh boundaries, while the SDK's didn't produce any. By that specifically I mean the corners of the inside of a box. I tried biasing the start position of a ray but this didn't help. Do anybody have some idea on how I can solve the 2 problems? The latter is relatively more important since I can use more ray samples anyway for the first question. Thanks a lot. Nicky

Share this post


Link to post
Share on other sites
Advertisement
Sign in to follow this  

  • Advertisement