Upper Bound on the Average Distortion of Multiple Description Coded Delay-Sensitive Sources

Multiple description coding (MDC) provides a means to combat erasure channels. Both rate and packet delay have significant contributions to end-to-end distortion of delay-sensitive sources. The previous works just considered exponential and fixed packet length distributions to evaluate the packet delay. In this letter, assuming a general (unknown) distribution with known average and variance of packet length, it is showed that average distortions of MDC and Single description coding (SDC) are upper bounded. Derived bounds are determined by encoding rate, packet arrival rate of network, acceptable delay threshold and first and second moments of packet length distribution.

[1]  Suhas N. Diggavi,et al.  Network Resource Allocation for Competing Multiple Description Transmissions , 2008, IEEE GLOBECOM 2008 - 2008 IEEE Global Telecommunications Conference.

[2]  Anthony Ephremides,et al.  Multiple description coding in networks with congestion problem , 2001, IEEE Trans. Inf. Theory.

[3]  Jonathan M. Pitts,et al.  Tight bounds for the tail of the packet waiting time distribution in buffered networks , 2001, Int. J. Commun. Syst..

[4]  Abbas El Gamal,et al.  Achievable rates for multiple descriptions , 1982, IEEE Trans. Inf. Theory.

[5]  Xinbing Wang,et al.  Optimal scheduling for multiple description video streams in wireless multihop networks , 2009, IEEE Communications Letters.

[6]  Vivek K. Goyal,et al.  Multiple description coding: compression meets the network , 2001, IEEE Signal Process. Mag..

[7]  L. Ozarow,et al.  On a source-coding problem with two channels and three receivers , 1980, The Bell System Technical Journal.