Share Email Print

Proceedings Paper

Integrated delay and error concealment for Internet voice applications with time-scale modification
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Packet delay and loss are two essential problems to the transmission of real-time voice over the best-effort Internet. Much effort has been involved in packet-level error control and delay jitter concealment. In our previous work, a time-scale modification of audio signal with per-packet adaptive playout algorithm is proposed to minimize the packet dropping at the receiver-end due to delay jitter. This work further extends the applicability of packet-based Synchronized OverLap-and-Add (SOLA) algorithm to an integrated delay/error concealment. Considering the timing relationship with the Forward Error Correction (FEC) parity-packet for loss recovery, the proposed adaptive playout algorithm estimates and adapts to packet loss as well as delay jitter. To enhance the playout quality, per-packet time-stretching factor is bounded by the content classification module, which classifies the audio signal into different categories. We also investigate the impact of stretching-ratio transition strategy to the perceived quality. To demonstrate the proposed adaptive playout, both analysis and performance evaluation is provided by comparing it to the silenced-based adaptive playout.

Paper Details

Date Published: 12 November 2001
PDF: 11 pages
Proc. SPIE 4518, Multimedia Systems and Applications IV, (12 November 2001); doi: 10.1117/12.448219
Show Author Affiliations
Fang Liu, Univ. of Southern California (United States)
JongWon Kim, Univ. of Southern California (South Korea)
C.-C. Jay Kuo, Univ. of Southern California (United States)

Published in SPIE Proceedings Vol. 4518:
Multimedia Systems and Applications IV
Andrew G. Tescher; Bhaskaran Vasudev; V. Michael Bove, Editor(s)

© SPIE. Terms of Use
Back to Top