Research On DASH-based Evaluation Model And Rate Adaptation Algorithm

In recent years,with the continuous growth of the business needs of the streaming media,the existing network bandwidth environment is far from enough to maintain all the video can be played in a high quality and no rebuffer.Therefore,how to effectively use the available bandwidth to maximize the resolution of the video buffer phenomenon and to improve the user's QoE streaming media business has become a hot research topic.The purpose of DASH(Dynamic Adaptive Streaming over HTTP)technology is to make full use of the bandwidth available in the network,and to bring more high-quality streaming media service quality experience to users.Therefore,it is very important to optimize and improve the adaptive algorithm in DASH technology,and to establish an effective and accurate evaluation model of DASH service QoE,which is of great significance to the development and improvement of DASH technology.In this paper,the factors affecting the QoE service DASH are tested and analyzed,and a BPNN based DASH service evaluation model is proposed based on the results of the analysis.Aiming at the problems existing in the current DASH rate adaptation algorithm,a new rate switching algorithm based on client buffer state is proposed.The main research points and innovations of this paper are as follows:(1)Aiming at the problem that the QoE evaluation model of traditional streaming media service is not applicable to the evaluation model of DASH service,a DASH service QoE evaluation model based on back propagation neural network(BPNN)is proposed.In this paper,first of all,a lot of subjective tests on the DASH business to obtain the average score of the user to play the video,and then determine the key factors that affect the DASH business through the two value variance method.Finally,based on the several key factors,through the BPNN of the sample to learn and constantly modify the weights to determine the impact of these key factors on the DASH business QoE weight.The experimental results show that the evaluation model can better predict the user's QoE based on the numerical value of several key factors when watching video,and the error of the prediction results and the actual value is within the acceptable range.(2)Aiming at the DASH technology to solve low initial video playback delay,improve the video quality and reduce the average video and buffer number and duration,reduce the switching frequency between different video quality problems,this paper proposes a rate adaptive algorithm based on QoE friendly.In this algorithm,through the design of a video clip with the lowest rate for the initial broadcast video file to reduce the initial delay;through the design of a fast algorithm to start as soon as possible to improve the video bit rate in the current bandwidth environment to meet the circumstances;the video is converted into multiple bit rate video,the the average network bandwidth is larger than the lowest bit rate video situation will not occur again video buffer phenomenon,to ensure the minimum initial delay conditions can improve the average video quality;by judging the average bandwidth of the current buffer state and a period of time to decide whether to switch the video bit rate,effectively reduce the average bit rate switching times.Experimental results show that the proposed algorithm can reduce the switching frequency between video and different quality video,and can effectively improve the user's DASH service QoE.Based on the DASH service QoE evaluation of BPNN model and the improved adaptive algorithm based on client buffer are starting from the key factors affecting the DASH business of QoE,the existing adaptive algorithm is improved.The research results of this paper have certain academic value and application value.