Recovery of voice packets using combined R-S erasure code with adaptive decoding and packet repetition concealment technique

It is well Reed-Solomon codes have very good burst error detection and correction capabilities. However, they are limited by their computation complexity and high decoding delay for recovering packet losses for delay-stringent application, like VoIP. Since the position of a lost packet is known, only error correction is required. So erasure codes can be used to recover lost packets. In this paper, Reed-Solomon (7,4) erasure code is employed to recover lost packets when three or fewer packets ore lost from a block of seven packets. The decoding is done adaptively and the unrecovered packets are replaced by using packet repetition concealment technique. It is found that the decoding delay and complexity is reduced when adaptive decoding is used. The performance of the various coding schemes are compared by the Perceptual Evaluation of Speech Quality (PESQ) Mean Opinion Score (MOS) values. The combined erasure code and packet repetition gives better performance.