Interactive Distributed Translucent Volume Rendering

Abstract

Translucent volume rendering is a robust and efficient direct volume-rendering technique for capturing optical effects, like subsurface scattering, translucency, and volumetric shadows. However, due to the limited computing and memory resources of the recent consumer graphics hardware, high-resolution volume data can still hardly be interactively visualized by this method. In this paper we present the theoretical aspects and implementation details of a parallelization scheme for translucent volume rendering. Our method is a three-pass parallel rendering algorithm with parallel compositing, based on object-space distribution of the data among the rendering nodes. In the first pass the 2D shadow maps are computed and sent to the effected nodes. In the second pass the nodes render their associated subvolumes by sequential translucent volume rendering. The generated framelets are then visualized by a dedicated display node in the third pass.