Research And Implementation Of Dynamic Adaptive Streaming Over HTTP

With the continuous improvement of our life quality, people’s demand forhigh-quality video and audio content is becoming more and more strong. And withbetter network facilities, it’s possible to transmit high-quality video and audio overInternet. For the past few years, with the burst popularity of smart phones and tablets,the way people connect to the Internet has changed a lot. Moreover, recently deployed3G and4G networks have provided people with a more convenient way to watchstreaming media content, which further increased the importance of streaming mediaservices in people’s lives. Under such situation, traditional streaming media technologycan no longer meet people’s needs, and a technology called Dynamic AdaptiveStreaming over HTTP (DASH) has become a new choice of researchers and streamingmedia service providers. However, there’s no criterion for evaluating rate adaptationalgorithms yet, and so far there isn’t much well recognized technical solutions in Chinaeither. Due to these reasons, research on DASH is very necessary.In this paper, we analyzed the feature of DASH and the MPEG-DASH standardand studied related technologies including video codecs, audio codecs and containerformat. After that, we studied some existing rate adaptation algorithms, analyzed therunning environment of the client and the goals of rate adaptation algorithm, whichhelped us to propose a buffer level base rate adaptation algorithm. On this basis, wedesigned and implemented a DASH system, and did further research on contentpreparing, the design and implementation of DASH client and network emulation.Finally, we designed several network scenarios to test our rate adaptation algorithm withthe system, experimental results show that the proposed rate adaptation algorithm canutilize bandwidth efficiently while providing smooth video quality.