Criterion-based speech packet loss recovery technique

Zhong Bo Li*, Sheng Hui Zhao, Jing Wang, Jing Ming Kuang

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

In order to solve the problem of speech packet loss on the Internet, a criterion-based packet-lovss recovery technique is proposed. It combines the sender-based forward error correction (FEC) mechanism and the receiver-based error concealment(EC) mechanism. The receiver-based EC method is simulated to get the virtual reconstruction for every packet and the criterion of minimizing mean square errors(MMSE) is adopted to obtain the redundancy information at the sender part. At the receiver part, the real reconstruction for the lost packet is obtained using the EC method together with the redundancy. Simulation results showed that the speech quality over the IP network is improved efficiently at a low bandwidth consumption with the proposed method.

Original languageEnglish
Pages (from-to)18-22
Number of pages5
JournalBeijing Ligong Daxue Xuebao/Transaction of Beijing Institute of Technology
Volume27
Issue numberSUPPL. 1
Publication statusPublished - May 2007

Keywords

  • Error concealment
  • Forward error correction
  • Speech packet loss

Fingerprint

Dive into the research topics of 'Criterion-based speech packet loss recovery technique'. Together they form a unique fingerprint.

Cite this