EUROGRAPHICS 2015 / O. Sorkine-Hornung and M. Wimmer (Guest Editors) Volume 34 (2015), Number 2 Partitioned Shadow Volumes J. Gerhards 1 , F. Mora 1 , L. Aveneau 2 , and D. Ghazanfarpour 1 1 University of Limoges - XLIM-CNRS, France 2 University of Poitiers - XLIM-CNRS, France {julien.gerhards, frederic.mora, lilian.aveneau, djamchid.ghazanfarpour}@xlim.fr Figure 1: Image rendered to a resolution of 1024 × 1024 with our novel shadow volume algorithm on a GTX 980 GPU. From left to right: Fairy (4.14ms, 174k triangles), Geodesic (2.27ms, 200k triangles), Raptor (8.96ms, 1 000k triangles). Abstract Real-time shadows remain a challenging problem in computer graphics. In this context, shadow algorithms gen- erally rely either on shadow mapping or shadow volumes. This paper rediscovers an old class of algorithms that build a binary space partition over the shadow volumes. For almost 20 years, such methods have received little attention as they have been considered lacking of both robustness and efficiency. We show that these issues can be overcome, leading to a simple and robust shadow algorithm. Hence we demonstrate that this kind of approach can reach a high level of performance. Our algorithm uses a new partitioning strategy which avoids any polygon clipping. It relies on a Ternary Object Partitioning tree, a new data structure used to find if an image point is shadowed. Our method works on a triangle soup and its memory footprint is fixed. Our experiments show that it is efficient and robust, including for finely tessellated models. Categories and Subject Descriptors (according to ACM CCS): I.3.7 [Computer Graphics]: Three-Dimensional Graphics and Realism—Color, shading, shadowing, and texture 1. Introduction Computing shadows is a problem almost as old as the com- puter graphics history. Light and shadow are two sides of the same coin, as both should be rendered with accuracy to cre- ate realistic pictures. This often comes down to a visibility problem. If the light source is not visible from a point, it is in shadow. Thus, computing hard shadows from a point light source is easier than computing soft shadows from an area light source. However, easier does not mean easy. Despite a large amount of works in the literature [ESAW11], com- puting high quality hard shadows in real-time remains chal- lenging for complex models. On graphic hardware, most of the methods rely on two different techniques: Shadow Maps and Shadow Volumes. If both are simple concepts, they are not easy to implement with robustness and efficiency. In this paper, we revisit the Shadow Volume BSP algorithm, a class submitted to EUROGRAPHICS 2015.