Quality of Experience Study on Dynamic Adaptive Streaming Based on HTTP

In this paper, the Quality of Experience (QoE) on Dynamic Adaptive Streaming based on HTTP (DASH) is researched. To study users' experience on DASH, extensive subjective tests are firstly designed and conducted, based on which, we research QoE enhancement in DASH and find that DASH ensures more fluent playback (less stall) than constant bitrate (CBR) streaming to promote users' satisfaction especially in mobile networks. Then we adopt two-way analysis of variance (ANOVA) tests in statistics to identify the effect of specific factors (segment bitrate, bitrate fluctuation pattern, and bitrate switching) that impair users' experience on DASH. The impairment functions are then derived for these influence factors based on the Primacy and Recency Effect, a psychological phenomenon that has been proved to exist in users' experience on DASH in this paper. And the final QoE evaluation model is proposed to provide high correlation assessment for QoE of DASH. The good performance of our QoE model is validated by the subjective tests. In addition, our QoE study on DASH is also applied for QoE management to propose a QoE-based bitrate adaptation strategy, which promotes users' experience on DASH more strongly than the strategy based on QoS.