Research On The User Behavior-Based QOE Evaluation Method For Http Streaming Service

Pacing into the information era, people’s demands for accessibility of multimedia materials increase significantly, which will inevitably lead to the emergence and development of streaming media technology. Nowadays, domestic and international video websites like Youtube, Hulu and Youku mainly provide video services for customers with the HTTP-based streaming media. Different from the RTSP/RTP-based streaming media, the HTTP-based streaming media service doesn’t require a specific server, and it works well on a common HTTP server. The prime purpose of the HTTP streaming media technology is to enhance users’satisfaction on the video services. Therefore, QoE (Quality of Experience) evaluation for the HTTP streaming media has become a hot topic this decade. Many academic and industrial investigators endeavor to improve the service quality and to optimize the network performance according to the evaluation results.This paper primarily focuses on the definition of QoE and its evaluation methods. Then we summarize the major progresses of HTTP mobile streaming media technology as well as the potential challenges associated with evaluation of the QoE. Furthermore, a summary is made after detailed investigations on1) the influencing factors and2) evaluation methods for the QoE of the HTTP streaming media.As indicated by previous studies, the QoE of HTTP streaming media is mainly controlled by three factors, i.e. the network layer, the application layer and the user layer. Because of the subjectivity and inaccessibility of data, the previous studies generally neglected impacts of user layer, and alternatively emphasized on the other two factors. However, for the variability of users’environments, this inclination would significantly lead to deviations of the evaluation results from users’ actual experiences. After the preliminary investigations, we analyze users’ behaviors during they viewing mobile videos, and put forward a new user behavior-based evaluation method for the QoE of HTTP streaming, which combines both users’behaviors and a series of objective parameters. According to users’action records, this method evaluates uses’ expectation for videos’fluency and clarity as well as their moods when viewing videos. After that, customers’satisfaction degrees for the videos are individually modified on the basis of the evaluation results.At last, according to the proposed evaluation method, we design and construct an Android mobile streaming media client, which is used to collect experimental data. Through analyzing these data, we validate that users’pause behaviors during video experiences actually affect the evaluations on QoE.