強健性寬頻語音編碼系統之研究

Abstract

在數位傳輸系統，訊源控制通道解碼的研究旨在補償通道錯誤所造成的解碼失真。訊源編碼輸出的量化索引序列存在殘餘冗息，若能在通道解碼過程妥善地利用這一項事前資訊，將可對接收端的解碼效能有所提昇。我們提出一個訊源控制通道解碼機制，利用寬頻語音編碼處理後所殘留的訊源冗息提昇其參數重建的強健性。有別於以往基於單一敘述的演算法，我們將引入多重敘述向量量化，並改良其解碼演算法以估測最大後驗機率。實驗模擬顯示在重建寬頻語音編碼參數上，訊源冗息與多重敘述傳輸機制的整合能有效提昇其解碼效能。

For digital speech communication, the system performance will degrade after compressed signals are transmitted over noisy channels. In order to combat channel errors, the signals will be protected by means of source-controlled channel decoding methods. When signals are quantized by the source encoder, quantizer indexes have residual redundancy due to non-uniform distribution and correlations between them. By taking advantage of such a priori knowledge, we propose a source-controlled channel decoding scheme which incorporates the source residual redundancy into the channel decoding process. Also proposed is a modified decoding algorithm based on the multiple description vector quantizer. Experiments on the wideband speech coding showed that the proposed multiple description transmission scheme outperforms single description scheme in parameter reconstruction.