Dynamic Real-time Soft Shadows

The fast development of fields of computer real-time rendering has attracted much attention. Presently, computer graphics hardware has evolved at a rapid rate: it is more than ten times faster (outstripping Moore's Law). Shadows are important to computer graphics because they add realism and help the viewer identify spatial relationships. Many shadow generation techniques developed over years have been used successfully in video games, movie productions and virtual reality. It is still challenging, however, to compute high-quality shadows in real-time for dynamic scenes.This thesis consists of three main parts. The first part is a short introduction for the research background and significance. Then, we present the main research contribution of this thesis. And the final is the overall organization of our thesis in details.The second part is a detailed description of shadow techniques, especially real-time soft shadows in dynamic scenes. Image rendering, real-time shadow algorithm and soft shadow algorithm are primary research contents in our thesis. Much techniques based on image rendering could be introduced into shadows, such as texture mapping, point rendering, antialiasing. Real-time shadow rendering techniques are the basis of generating soft shadows. In this time shadow maps and shadow volumes technique become two branches of real-time shadow rendering. Shadow maps are computed in image space, rendering shadow with depth maps' information in scenes. Comparison to shadow maps, instead of computing in image space, shadow volumes generate shadows based on object geometry information in object space. The character of image space algorithm is ensuring real-time rendering but having the poor-quality performance, while object space algorithm brings high-quality image but low frame rates.The last part we present a hybrid algorithm for rendering soft shadows accurately and efficiently. Our approach is related to shadow maps and provides associated benefits. The computation time and memory requirements for adding soft shadows to complexity. Layered attenuation maps achieves interactive rendering rates, but limits sampling flexibility, while coherence-based raytracing of depth images is not interactive, but removes the limitations on sampling and yields high quality images at a fraction of the cost of conventional raytracers. Combining the two algorithms allows for rapid previewing followed by efficient high-quality rendering. We choose the two image-based rendering techniques because area light sources are localized in space, soft shadow computations are particularly well suited.