I'm implementing bidirectional path tracing. I want to simulate perfect refraction. I read in the Veach thesis that a perfect refraction BSDF is non-symmetric. Does it mean that I need to use different computations for refraction BSDF (BTDF) for eye subpaths and light subpaths? It seems strange for me, as BPT differs from PT only in the way I generate the path and throughput computations should be the same, as in PT all BSDFs are computed in the same way. Is that non symmetry relevant only for particle based techniques like Photon Mapping?
Asked
Active
Viewed 139 times
