Abstract: While traditional graphics techniques provide for the realistic display of three-dimensional objects, these methods often lack the flexibility to emulate expressive effects found in the works of artists such as Michelangelo and Cezanne. We introduce a technique for capturing custom artistic shading models from sampled art work. Our goal is to allow users to easily generate shading models which give the impression of light, depth, and material properties as accomplished by artists. Our system provides real-time feedback to immediately illustrate aesthetic choices in shading model design, and to assist the user in the exploration of novel viewpoints. We describe rendering algorithms which are easily incorporated into existing shaders, making non-photorealistic rendering of materials such as skin, metal, or even painted objects fast and simple. The flexibility of these methods for generating shading models enables users to portray a large range of materials as well as to capture the look and feel of a work of art. (Color images can be found at http://www.cs.utah.edu/npr/papers/LitSphere_HTML.)