This research involves, (1) developing techniques for the design of optimal residual vector quantizers (RVQs), and (2) characterizing the performance of the resulting RVQs on a variety of synthetic and natural information sources. In particular, a new formulation of the optimization problem for RVQ design is researched. This will lead to an understanding of the necessary conditions for optimality and design of residual vector quantizers for a number of scalar sources with a decrease in distortion on the order of 10-15 dB.
