Vector quantization of spherically invariant random processes

Vector quantization of spherically invariant random processes (SIRP) is considered. Especially, trellis coded quantization (TCQ) and lattice vector quantization (LVQ) are investigated. For performance evaluations a random number generator has been developed producing sequences which can be regarded as SIRP realizations. It turns out that in most cases the TCQ outperforms all other investigated quantization methods, even those LVQ schemes which are matched to the properties of SIRP sources. Comparisons with bounds from rate distortion theory are given as well.