Multiple Description Error Mitigation Techniques for Streaming Compressed Audio Over a 802.11 Wireless Network

Cheng CI, Jiang W. Multiple Description Error Mitigation Techniques for Streaming Compressed Audio Over a 802.11 Wireless Network. In: AES Convention 120. Audio Engineering Society; 2006. Paper 6663. Available from: https://aes.org/publications/elibrary-page/?id=13467

@inproceedings{Cheng2006_13467,
  author = {Cheng, Corey I. and Jiang, Wenyu},
  title = {{Multiple Description Error Mitigation Techniques for Streaming Compressed Audio Over a 802.11 Wireless Network}},
  booktitle = {AES Convention 120},
  note = {Paper 6663},
  year = {2006},
  month = may,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=13467}
}

TY  - CPAPER
TI  - Multiple Description Error Mitigation Techniques for Streaming Compressed Audio Over a 802.11 Wireless Network
AU  - Cheng, Corey I.
AU  - Jiang, Wenyu
T2  - AES Convention 120
M1  - Paper 6663
PY  - 2006
DA  - 2006/05/06
UR  - https://aes.org/publications/elibrary-page/?id=13467
PB  - Audio Engineering Society
LA  - en
AB  - This paper presents several multiple description (MD) coding techniques for error mitigation of compressed audio streamed over an 802.11b/g wireless network. Loosely speaking, an MD encoder generates several descriptions of the same source, and an MD decoder recreates the best estimate of the source from the set of descriptions it successfully receives. We propose a design for an MD architecture and simulate its integration into the AAC codec. We use packet loss traces gathered from an actual 802.11 b/g network to simulate the proposed codec’s error mitigation properties for various network traffic conditions. We examine how tuning several of the proposed codec’s parameters would affect the sound quality and overall bitrate of the proposed codec. Specifically, we show how interleaving, renormalization, and low-frequency variance estimation techniques can be used in conjunction with hierarchical correlating transforms to improve the sound quality of multiple description codecs.
ER  -