On Distributed Quantization in Scalable and Predictive Coding

The study of distributed coding system design for correlated sources has largely focused on the basic problem, with the implicit assumption that practically motivated extensions involving scalable or predictive coding may be obtained in a straightforward manner. In this paper, we show that naively extended approaches yield poor ratedistortion performance. In fact, inherent conflicts arise between distributed quantization and scalable or predictive coding. Distributed predictive coding is further plagued by design instability of closed loop predictors, whose impact has been recognized in the context of single source coding, but is greatly exacerbated in the case of distributed coding. We propose a general framework that allows for and controls mismatch between encoder and decoder estimates in the base-layer or prediction loop, respectively, in distributed scalable or predictive coding. Simulation results show substantial gains over single source (separate) scalable or predictive coding techniques, as well as over naive extensions to incorporate scalability or prediction in distributed coding.