Posted 29 October 2012 - 11:17 AM
Depending on what you want to do, it might just be quicker to write your own volume renderer, assuming you already have a simple shader based renderer. If you have access to compute shaders, it is trivial to setup a shader which finds intersections in between your volume box / primitive and your camera rays, and step through each voxel in between those points. Or you can raterize a box with the correct UVs for your voxels both front and back face, and then use the resulting rts to know the exact start and end uvs for each ray. More work to get going, but no math and might be faster, depending on your volume primitives.