Quality of Experience Assessment for HTTP Based Adaptive Video Streaming

Abstract

The expected optimal outcome of users' video quality requirements should be the best Quality of Experience (QoE). Integrating and evaluating QoE in a video streaming system is a rather complex process that requires a comprehensive design of subjective databases and video quality metrics. The goal of this paper is to examine the parameters that affect QoE in adaptive video streaming, as well as metrics for evaluating video quality. The paper reviews and analyses 21 QoE metrics, out of which 4 are general video evaluation metrics, 7 are adaptive video streaming evaluation metrics, and 10 are hybrid metrics and their usability for QoE measurement for HTTP-based adaptive streaming (HAS). The LIVE-NFLX-II database is used to compare the metrics. The hybrid metrics that combine one of the full reference metrics with the no-reference metric that uses a combination of average bitrate, rebuffering percentage, and average bitrate switching magnitude, has yielded the best correlations with the subjective quality ratings.

Keywords

• hybrid metrics
• http adaptive streaming
• live-nflx-ii database
• quality of experience
• video quality metrics