Rate-distortion function for nonstationary time-varying autoregressive (TVAR) processes

Abstract

The Gaussian TVA.R process is of primary interest because of its wide applicability to the modeling real-world sources, e.g., speech signals and rasterized image signals. We then investigated and constructed the rate-distortion function R ( D ) for the class of time-varying nonstationary Gaussian signals. The numerical evaluations of the R ( D ) function for 1-D Gaussian TVAR processes are also presented.