Abstract | ||
---|---|---|
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. |
Year | DOI | Venue |
---|---|---|
2010 | 10.1111/j.1467-8659.2010.01737.x | Comput. Graph. Forum |
Keywords | Field | DocType |
theoretical optimal discretization,octahedron normal vector,discretization error,common sphere parameterizations,normal vector representation,floating-point normal vector,single precision floating-point normal,controllable error,floating point | Single-precision floating-point format,Discretization,Discretization error,Computer science,Floating point,Algorithm,Theoretical computer science,Minifloat,Normal | Journal |
Volume | Issue | ISSN |
29 | 4 | 0167-7055 |
Citations | PageRank | References |
20 | 0.84 | 4 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Quirin Meyer | 1 | 63 | 4.80 |
Jochen Süßmuth | 2 | 138 | 11.30 |
Gerd Sußner | 3 | 23 | 2.24 |
Marc Stamminger | 4 | 1465 | 112.74 |
Günther Greiner | 5 | 598 | 80.74 |