Research On Network Streaming Media System Based On Scalable Video Coding Technology

Video information is more often from the network transmission along with the continuous development of Internet and multimedia technology. Network streaming media technology came into being, and has made great strides. However, the diversity of user requirements, the network's inherent properties (heterogeneity, bandwidth fluctuation and unreliability) and the heterogeneous of video terminals limits the development of streaming media technology. Bandwidth fluctuation determines it is difficult for streaming media technology to provide real-time Streaming media services for different clients. Video compression coding can no longer be only for the storage, but also need to adapt to the complex network environment, while ensuring the efficiency of high compression, it requires that the video stream is scalable so that it can adapt to network bandwidth fluctuations. In response to this mandate, scalable video coding technology came into being.Scalable video coding technology requires that we can extract code stream with different spatial resolution, time resolution and bitrates from encoded video stream by a simple packet throwing, stream truncating or other operations. It requires encoded video stream can provide spatial scalability, time scalability and quality scalability. In the October 2004, MPEG meeting decides to use H.264-based SVC method proposed by HHI of Germany, and proposed a software reference model. In January 2005, ISO's MPEG and ITU-T's VCEG agreed to make SVC be an extension set of H.264/AVC and decide that the draft proposed by the Joint Video Team, included it in Appendix G for the H.264/AVC. After that each JVT meeting is about revising and improving of SVC technology draft.This paper mainly researches on scalable video coding technology based on H.264/AVC and Network bandwidth measurement technology, designs and implements a network streaming media system based on scalable video coding technology. In order to achieve the purposes of full use of network bandwidth and client-side's real-time playing. This system can detect the current network available bandwidth, and select a code stream with different spatial layer, time layer and quality layer based on the current network conditions and transmit it to the client. In order to verify the correctness of the system, we build a wide-area network environment, use TC tool Under Linux to change network available bandwidth, and verify quality scalability, spatial scalability and time scalability respectively under different network available bandwidth.