Hardware accelerated rendering of CSG and transparency

摘要

This paper describes algorithms for implementing accurate rendering of CSG and transparency in a hardware 3D accelerator. The algorithms are based on a hardware architecture which performs front-to-back Z-sorted shading; a multiple-pass algorithm which allows an unlimited number of Z-sorted object layers is also described. The multiple-pass algorithm has been combined with an image partitioning algorithm to improve efficiency, and to improve performance of the resulting hardware implementation.