Meyer, QuirinSuessmuth, JochenSussner, GerdStamminger, MarcGreiner, Guenther2015-02-232015-02-2320101467-8659https://doi.org/10.1111/j.1467-8659.2010.01737.xIn this paper we analyze normal vector representations. We derive the error of the most widely used representation, namely 3D floating-point normal vectors. Based on this analysis, we show that, in theory, the discretization error inherent to single precision floating-point normals can be achieved by 250.2 uniformly distributed normals, addressable by 51 bits. We review common sphere parameterizations and show that octahedron normal vectors perform best: they are fast and stable to compute, have a controllable error, and require only 1 bit more than the theoretical optimal discretization with the same error.10.1111/j.1467-8659.2010.01737.x1405-1409