Research On Streaming Media Video Quality Adaptation Technologies Based On SVC-P2P

With the rapid development of mobile Internet and triple play, P2P streamingmedia system will inevitably provide services in heterogeneous environment. ScalableVideo Coding (SVC) can offer three dimensions in scalability for a video file: temporalscalability, spatial scalability, and quality scalability. The application of Scalable VideoCoding (SVC) in P2P streaming media system provides new ideas for video qualityadaptation. Therefore, this paper focuses on the research of video quality adaptationtechnologies in P2P streaming media system.Firstly, data characteristics of SVC videos are analyzed,general video qualityadaptation technologies of P2P streaming are introduced, and the key problems of videoquality adaptation based on SVC in heterogeneous environment are cleared.Secondly, video quality adaptation is aimed at a better view experience to user,criteria for user’s experience is particularly important. Most of the current researchesuse QoS or the downloaded video layers as the criteria to evaluate video quality.However, these criteria can’t fully reflect whether the video quality meets user’s needs.In this paper, from the perspective of the user, QoE (Quality of Experience) is chosen asevaluation standard of user’s experience. Through comprehensive analysis of keyparameters that affect on QoE, such as coding parameters, network parameters, videotype, and user terminal and so on, this paper gives a QoE prediction model based onCART-Adaboost algorithm. The simulation results show that, compared with otherprediction methods, the prediction model in this paper has better prediction accuracyand real-time performance.Lastly, in order to adapt to network dynamics in real time, and to provide user withbetter video quality experience, the following problems of video quality adaptationalgorithm still need to be solved: when and how to adjust video quality.The generalmethods are that according to the change of network bandwidth, video quality real-timeadjustment is achieved by increasing or decreasing the downloaded temporal, qualityenhancement layers, or the bit rate. However, experimental analysis indicates thatdifferent video types, different coding parameters setting and different layercombination contain different data size, and then user have different sensitivity inchanges of video quality caused by different dimensions. All of quality adaptationalgorithms above don’t work well. Based on QoE prediction model, this paper gives a video quality adaptation algorithm, which is according to the change of networkbandwidth and predicted results by QoE model to determine whether to adjust videoquality, and to select the current the optimal combination by using the QoE model. Thesimulations reveal that our algorithm can not only adapt to network change in real time,but also select the optimal combination in current condition for heterogeneous terminalsto maximize quality of experience.