Rendering geometric models with complex surface materials in arbitrary lighting environments is a challenging problem. In order to relight and render geometries covered with complex, measured BTFs two problems have to be addressed: The memory problem resulting from the large size of the measured BTF data and the light integration problem resulting from summing up the contributions from all measured lightsources. In this paper we describe how highly efficient BTF compression methods like Local-PCA and suitable representations of environmental light based on Spherical Harmonics can be combined leading to fast environmental lighting for efficiently encoded BTFs. As a side effect the method supports Precomputed Radiance Transfer.