We describe a technique for real-time rendering of dynamic, spatially-varying BRDFs in static scenes with all-frequency shadows from environmental and point lights. The 6D SVBRDF is rep-
resented with a general microfacet model and spherical lobes fit to its 4D spatially-varying normal distribution function (SVNDF). A sum of spherical Gaussians (SGs) provides an accurate approximation with a small number of lobes. Parametric BRDFs are fit on-the-fly using simple analytic expressions; measured BRDFs are fit as a preprocess using nonlinear optimization. Our BRDF representation is compact, allows detailed textures, is closed under products and rotations, and supports reflectance of arbitrarily high specularity. At run-time, SGs representing the NDF are warped to align the half-angle vector to the lighting direction and multiplied by the microfacet shadowing and Fresnel factors. This yields the relevant 2D view slice on-the-fly at each pixel, still represented in the SG basis. We account for macro-scale shadowing using a new, nonlinear visibility representation based on spherical signed distance functions (SSDFs). SSDFs allow per-pixel interpolation of high-frequency visibility without ghosting and can be multiplied by the BRDF and lighting efficiently on the GPU.