A forward-backward voice packet loss concealment algorithm for multimedia over IP network services

Abstract

In this paper, we propose a voice packet loss concealment algorithm in order to improve voice quality for both multimedia over IP and voice over IP services. The proposed algorithm estimates the coding parameters of lost frames by combining forward and backward prediction from the good frames before and after the lost frames. The performance of the proposed algorithm is evaluated on the ITU-T G.729 coder, and it is compared with the performance of the conventional algorithms in terms of objective and subjective quality measures. From the PESQ score comparison and the listening test, it is found that the proposed algorithm provides better voice quality than the conventional ones.