A Study of QoE-Aware Adaptation Mechanism for DASH Video Streaming based on Objective Visual Quality Assessment

Abstract

Dynamic Adaptive Streaming over HTTP (DASH) is a largely used video streaming technique. One key point is its adaptation mechanism which resides at the client’s side. This allows various context-aware adaptation strategies in order to optimize the overall Quality of Experience (QoE) of the video streaming. In this paper, we present a study on an adaptation mechanism which uses an objective visual quality assessment, namely the Structural Similarity Index Measurement (SSIM) metric, as a key criterion for adaptation. More specifically, the SSIM helps to maximize the effective use of the available bandwidth, in the sense that we adopt a higher bitrate not only because it is allowed by network conditions, but also because it does bring a significant visual quality improvement (measured through SSIM metric). In this way, an upgrade in bandwidth consumption will be allowed only if there is a real contribution to visual quality. This study has been tested through a series of experimental results obtained with several strategies for the choice of the threshold value. Our tests are all based on real mobile-network traffic traces and real video sequences.