Optimal Biorthogonal Filter Banks for Multiple Description Coding

Pier Luigi Dragotti1 and Martin Vetterli1;2 1 Dept. of Electrical Engineering, Swiss Federal Institute of Technology CH-1015 Lausanne, Switzerland 2 EECS Dept., UC Berkeley CA 947020 e-mail: fdragotti,[email protected] http://lcavwww.epfl.ch Abstract | Multiple description coding (MDC) is a source coding technique for information transmission over unreliable networks. In MDC, the coder generates several different descriptions of the same signal and the decoder can produce a useful reconstruction of the source with any received subset of these descriptions. This paper addresses the design of a two-channel biorthogonal filter bank for optimal MDC of wide sense stationary Gaussian input processes. The problem is solved in the frequency domain, the redundancy rate-distortion curve is constructed and, for each given redundancy, filter responses are designed to meet the corresponding point in the redundancy rate-distortion curve. An application of the proposed algorithm to a first order GaussMarkov process is presented and for this case also approximated FIR solutions are proposed.