Enhanced forward error correction scheme in ACELP based codec for voice over IP network services

Abstract

Forward Error Correction (FEC) is a packet loss concealment (PLC) mechanism used to recover lost packets via the transmission of a set of redundant packets along with the original ones. Given its performance in improving the end-to-end error recovery capability of lossy networks as well as maximizing the end-to-end quality of service (QoS), FEC is usually considered to be an interesting choice for real-time applications like voice over Internet protocol (VoIP). Note that this mechanism has a low efficiency against consecutive packet losses, which heavily affects the perceptual speech quality in a VoIP environment and could even produce a total desynchronization between the coder and the decoder. The objective of this paper is to introduce a new FEC scheme that can address the aforementioned issue. We use also a Time Scale Modification (TSM) procedure that enhance the performance of our proposed FEC scheme in order to further improve the performance of our scheme under severe network conditions. Subjective and objective evaluation metrics are used to evaluate the efficiency of the proposed method. Results show that the new FEC method achieves a Perceptual evaluation of speech quality (PESQ) value exceeding 3.5 at a 10% frame erasure rate with a noticeable enhancement in voice perception.