Meyer Q.,Computer Graphics Group Erlangen |
Sussmuth J.,Computer Graphics Group Erlangen |
Sussner G.,Realtime Technology |
Stamminger M.,Computer Graphics Group Erlangen |
Greiner G.,Computer Graphics Group Erlangen
Computer Graphics Forum | Year: 2010
In 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. © 2010 The Eurographics Association and Blackwell Publishing Ltd.