Research On The Channel Coding Technology, Transmission Mechanism, And QoE Evaluation Model For Adaptive Streaming

The channel encoding technology, the transmission mechanism, and the QoE (Quality of Experience) evaluation model for the HTTP adaptive streaming are researched in this thesis. The main contents comprise the following three parts:research on Unequal Error Protection (UEP) channel coding aimed at adaptive streaming, research on the adaptive streaming transmission mechanism based on the above UEP channel coding scheme, and research on the QoE evaluation model for the streaming services including constant bitrate (CBR) service and adaptive bitrate service.The purpose for the evolution of adaptive streaming, such as DASH (Dynamic Adaptive Streaming over HTTP), is to provide better experience on the media streaming service for the user. Therefore, it is necessary to optimize and improve the technical scheme of DASH, and to analyze and evaluate the users’ subjective perception on the DASH service, which can help to improve the technical scheme for DASH. In this thesis, the basic technical characteristics of the adaptive streaming (DASH) have been introduced first, and the primary drawbacks of DASH and the probable solutions are further discussed. Then to overcome the drawbacks of DASH and based on the scalable source encoding thought, a UEP rateless code scheme is proposed, and based on which, a novel adaptive streaming transmission mechanism is designed. Besides, the thesis also models the QoE evaluation algorithms for both CBR service and DASH service, and applies the above QoE research to manage and optimize the technology of DASH.The research work of this thesis covers the following 3 innovation points:1. An application layer channel encoding scheme with UEP charactristics, named as UEP LT Code, is proposed. The scheme provides protection measures with different intensities to data of different important level. The theoretical analysis and simulation results prove that the proposed UEP LT Code scheme can ensure the source data with different important level have the high diversity on decoding probability and decoding priority with a very small overhead on the decoder.2. Based on the above UEP LT Code scheme and the H.264/SVC source scalable video encoding scheme, a novel adaptive streaming transmission model HASUR (HTTP Adaptive Streaming based on UEP Rateless Code) is proposed. The HASUR model not only overcomes the shortage of DASH by saving the storage resource of the server and with no need to estimate the bandwidth in advance, but also guarantees the great self-adaptation for the stream to the chanagable network environment. Besides, the model also provides good performance on efficient resilience against packet loss because of the UEP property of the channel coding scheme. Compared with DASH and HLS (HTTP Live Streaming) services, the HASUR model can provide more fluent and much clearer playback experience to the user under the end to end scenario.3. The QoE of adaptive streaming service is researched. The QoE promotion of the DASH service compared with the CBR service is studied. And then two QoE objective assessment models are proposed to evaluate the users’subjective perception on the CBR services and DASH services respectively. Based on the above QoE research, a QoE-driven bandwith adaptation strategy is designed to improve the performance of DASH service.