Speech Quality Degradation in Packet Loss Environment at Specific Speech Class
Minki Lee, Sung-Wan Youn, Kyung‐Tae Kim, Hong-Goo Kang · ITC-CSCC :International Technical Conference on Circuits Systems, Computers and Communications · 2007
In a specific context of voice over packet networks, one or several encoded speech frames are grouped into a single packet, which represents a data block on the network. Even the effort to control the variability in IP networks does not guarantee the quality of service (QoS) in the IP networks. Thus, it is necessary to deal with the audible artifact caused by packet losses. In this paper, we show experiment results to observe the impact of error propagation to speech quality in predictive coders, then we analyze the impact of the packet loss to error propagation in a low bit-rate speech codec. We choose the G.729 as a low bit-rate standard coder which operates at 8 kbit/s and encodes the speech signal based on CELP algorithms. Moreover, we investigate the relationship of packet loss to the speech characteristics and analyze which parameter plays an important role to each speech class.