Demonstration of adjusting forward error correction with quality scaling for TCP-friendly streaming MPEG

Huahui Wu, Mark Claypool, Robert Kinicki
2004 Proceedings of the 12th annual ACM international conference on Multimedia - MULTIMEDIA '04  
OVERVIEW The growth in the power and connectivity of the Internet has sparked an even larger growth in streaming media. The sheer number of possible users and applications at any point in time raises the probability of streaming multimedia flows encountering congestion. To overcome short-term congestion and avoid long-term congestion collapse, there is a growing consensus that Internet applications must be TCP-Friendly, with proposed approaches to detect and punish non-TCP friendly flows.
more » ... TCP, new TCP-friendly streaming media protocols refrain from retransmissions to avoid delay and jitter, but they are susceptible to quality degradation from packet loss. While multimedia applications can tolerate some data loss, excessive packet loss during congestion yields unacceptable media quality. Since video encoding involves interframe dependencies to achieve high compression rates, the random dropping of packets by routers can seriously degrade video quality. For example, as little as 3% MPEG packet loss can cause 30% of the frames to be undecodable. Streaming media flows often utilize lower latency repair approaches, such as Forward Error Correction (FEC), in conjunction with TCP-Friendly protocols to deliver streaming applications over the Internet. However, FEC requires redundant repair data to be added to the original video stream. Current approaches use either apriori, static FEC choices or adapt FEC to perceived packet loss on the network without regard to TCP-Friendly data rate constraints. When a streaming video operates within TCP-Friendly bitrate limits, adding FEC will reduce the effective transmission rate of the original video content. To preserve real-time streaming media playout, multimedia servers must scale back their streaming data rate to match the TCP-Friendly data rate using media scaling. With quality scaling, a widely used form of media scaling, the multimedia server adjusts the quantization level before transmission. A multimedia application can chose to increase the quantization level to save capacity for the FEC overhead. Hence, selecting the optimal amount of FEC and the optimal quantization level can be cast as a constrained optimization problem that attempts to optimize the quality of the video stream. APPROACH When the quantization level lq increases, the frame size and video quality decrease. Previous research shows the bitrate of a MPEG stream can be approximated by an exponential function of lq and our experiments suggest frame size can also be estimated by an exponential function of quantization level. We use the VQM metric developed by the Institute for Telecommunication Sciences 1 as an objective video quality measurement tool. VQM has a high correlation with subjective video quality assessment and has been adopted by ANSI as an objective video quality standard. The VQM tool takes an original video and a distorted video as input and returns a distortion value D between 0 (no distortion) and 1 (highest distortion). We encode the video with different quantization levels and use the VQM tool to measure the distortion. Our results show the distortion D could be also approximated by an exponential function of the quantization level. When FEC is used in the presence of packet loss, the successful frame transmission probability is:
doi:10.1145/1027527.1027564 dblp:conf/mm/WuCK04 fatcat:5pb5ssfvwbhrre6zieuc2dbcwy