TNT logo
   

Video streaming

Dipl.-Ing. Julia Schmidt
M.Sc. Zhijie Zhao
Show all

Background

With growing number of the high bit rate internet access, video streaming over IP network has become more important recently. Nevertheless, the IP network is a best-effort packet-switched network, and therefore the video transmission over IP network must cope with problems such as time-varying bandwidth, delay, jitter and packet losses. On the one hand, the present video coding standards have insufficient scalability to provide a wide range of sending data rate. The server must have a set of bit streams for the same video content to be able to serve different clients with different available bandwidth or to serve a client in case of congestion in the network. On the other hand, the present video coding standards achieve the high coding efficiency by exploiting the correlation in video signals. It results the dependency between the packets of a bit stream and causes error drift or interrupt in decoding process by packet loss. In an environment with high packet loss rate like wireless environment, the retransmission of lost packets causes latency and loads the congested network. Therefore, the error resilience and concealment are very important in the streaming system

Goal

The main goal of this project is to build-up an uni-cast streaming system using scalable extension of H.264/AVC. This streaming system should deliver a good performance in case of intense bandwidth fluctuation as well as in case of high packet loss rate.

Approach

To avoid a large number of bit streams for the same video content at the server side, the scalable extension of H.264/AVC is used as video format for the video streaming system in this project. On the server, a congestion control algorithm based on bandwidth estimation is applied. A new error concealment concept for the base and enhancement layers of a scalable bit stream will cope with packet loss.

References

Recent Publications

Show all publications
Zhijie Zhao, Alex Giladi, Joern Ostermann
Proposal for MPEG-2 Transport Stream Extensions for DASH
ISO/IEC JTC1/SC29/WG11 MPEG2010/M20119, March 2011
Jong-Seok Lee, Francesca De Simone, Naeem Ramzan, Zhijie Zhao, Engin Kurutepe, Thomas Sikora, Joern Ostermann, Ebroul Izquierdo, Tourad Ebrahimi
Subjective evaluation of scalable video coding for content distribution
ACM International Conference on Multimedia, pp. 65-72, Firenze, Italy, October 2010
Zhijie Zhao, Joern Ostermann
Video Streaming Using Standard-Compatible Scalable Multiple Description Coding Based on SVC
IEEE International Conference on Image Processing, Hong Kong, September 2010