Psychometric analysis of the effect of buffering control on user-level QoS in an interactive audio--visual application

  • Authors:
  • Yoshihiro Ito;Shuji Tasaka;Yoshihiko Fukuta

  • Affiliations:
  • Institute for Clarity in Documentation, Wallamaloo, New Zealand;Institute for Clarity in Documentation, Wallamaloo, New Zealand;Institute for Clarity in Documentation, Wallamaloo, New Zealand

  • Venue:
  • Proceedings of the 2004 ACM workshop on Next-generation residential broadband challenges
  • Year:
  • 2004

Quantified Score

Hi-index 0.01

Visualization

Abstract

This paper studies subjective tradeoff between fidelity and latency in an interactive audio--visual application over the Internet. In audio--video transmission, its temporal structure is disturbed by delay jitter of packets. The fidelity means how exactly the temporal structure is preserved. The degradation in the fidelity can be improved by playout buffer. Longer buffering time can suppress the disturbance caused by larger delay jitter, while it increases the latency, which is end--to--end delay, that is, latency. the difference between the time when media are generated at the sender and the time when the media are output at the receiver. The increase of the latency degrades the interactivity between users. The decrease of the fidelity and the increase of the latency degrade user--level QoS in interactive audio--visual applications. However, by changing the buffering time, we can improve either the fidelity or the latency. This means that the buffering control causes subjective tradeoff between the fidelity and the latency. Therefore, we must investigate a way of setting appropriate buffering time. In this paper, we study the effect of buffering time on user--level QoS in an interactive audio--visual application. In order to assess the user--level QoS quantitatively, we adopt psychometric methods. By experiment, we find an appropriate value of the initial buffering time in our experimental environment.