Codage conjoint source-canal stochastique: Application au codage robuste de la parole a bas debit

In this paper, we propose a stochastic joint source-channel scheme developed for robust encoding of spectral speech LSF parameters. The system, named LSF-SSCOVQ-CR encoder, is an LSF encoding scheme based on a Reduced Complexity Stochastic Channel Optimized Vector Quantization (SCOVQ-RC) scheme, modified according to the Split VQ (SVQ) concept. For transmissions over noisy channel, we will show first that the SCOVQ-RC encoder presents many advantages over the conventional COVQ. After that, we applied our LSF-SSCOVQ-CR encoder (with weighted distance) for the robust encoding of LSF parameters of the 2.4 Kbits/s MELP speech coder operating over a noisy channel. The simulation results will show that the proposed LSF encoder, incorporated in the MELP, ensure better performances than the original MELP MSVQ of 25 bits/frame; especially when the transmission channel is highly disturbed.