Publikationen

Creating photo realistic images from real word complex materials is a great challenge: the reflectance function of materials, especially fabrics, has glossy or specular highlights, reflectance anisotropy and retroreflections. This increases greatly the complexity of rendering. Bidirectional Texture Functions (BTFs), i.e. 2D textures acquired under varying illumination and viewing directions, have been used to render complex materials. However, the acquisition of textures for BTF requires up to now expensive setups and the measurement process is very timeconsuming as the directional dependent parameters (lighting and viewing directions) have to be controlled accurately. This paper will present in detail a new low cost programmable device for the rapid acquisition of BTF datasets. The device allows to acquire BTF databases at a fraction of the cost of available setups, and allows to experiment when a texture resolution and sample density increase in the parameter space is not perceivable by an observer of the renderings. The paper proves that using smaller resolution textures and decreasing the samples in parameter space does not lead to a loss of picture quality.