Wyner-Ziv coding of three correlated Gaussian sources using punctured turbo codes

In this article, we address the problem of distributed coding of three correlated memoryless Gaussian sources using punctured turbo codes. We first revisit the problem of distributed coding of two Gaussian sources. The impact of the distortion induced by the quantization of the side information on the performance of Wyner-Ziv decoding of two continuous valued Gaussian correlated sources is analyzed. Theoretic bounds as well as practical coding performances, when the side information is quantized, are given. This leads to the derivation of appropriate settings for the problem of distributed coding of three sources. The Wyner-Ziv theorem is then extended for three sources. Practical solutions using punctured turbo codes are described. Simulation results are presented and analyzed for different amounts of correlation