Weighted multi-pass method based on stochastic iteration random walk

Abstract

Multipass algorithms implement different rendering methods and combine their results. If carefully implemented, these algorithms can keep the advantages of individual methods. Algorithms of multipass global illumination may handle disjoint parts of light transport paths in different passes, and include all light transport only once in the final solution. On the other hand, they may also generate the same paths, thus their contribution should be weighted in order to get an unbiased estimation. In this paper a weighted combination of global ray bundle iteration and path tracing is presented. Heuristics for the weights are derived to get the benefits from both approaches. Results show significant improvements compared to both ray bundle iteration and path tracing.