Efficient and realistic shadow rendering using graphics hardware

Shadows have a great influence on the looks of a computer generated image. They are not only necessary for an authentic illumination of the scene, they also give cues about the position, size and shape of an object. The generation of shadows requires solving the problem of visibility, i.e. deciding whether the light source is visible from a point in the scene or not. Established methods like raytracing or radiosity are capable of rendering shadows, but are not suitable for real-time applications like games. Algorithms for real-time shadowing have been presented in the past, the two most popular and common being shadow volumes and shadow maps. Based on these fundamental algorithms many proposals have been made to improve them in terms of performance and appearance, because both algorithms have their drawbacks in either way. The recent appearance of programmable graphics hardware lead to completely new extensions of the existing algorithms previously not possible, especially hybrid algorithms that combine strengths or reduce drawbacks of shadow maps and shadow volumes. Two of those hybrid algorithms, shadow volume reconstruction and shadow silhouette maps, are described in this thesis along with possible implementations, with a focus on how they can benefit from graphics hardware