Vector quantizer.

摘要

The quantiser comprises an output vector code-table for partitioning a K-dimensional Euclidean signal space including an input vector consisting of K input samples (K larger than 1) to provide subspaces having min. distorted partitions w.r.t. the probability density of the input vectors and for storing the representative output vectors of the subspaces. A subtractor calculates the difference between the output vectors successively read from the code-table and the input vectors w.r.t. their individual components. The input signals are received in block form and the quantiser operates to determine an output vector having the least distortion w.r.t. an input vector. A corrector circuit converts each component of the output from the subtractor into an absolute value for determining the distortion of each element. Data compression can be accomplished by transmitting or recording a coded output including fewer sets of output vectors.