Quality of Experience Driven Rate Adaptation for Adaptive HTTP Streaming

A quality of experience (QoE) driven rate adaptation approach is proposed in this paper for adaptive HTTP streaming which jointly considers both bandwidth savings and video quality adaptation into account for the rate adjustment, beneficial to both service providers and subscribers. At each decision epoch, the adaptation algorithm accumulates for the time varying QoE of a viewer by accounting for all the impairments, namely, initial delay, quality transitions, and playback interruptions, which commonly occur during the playback of a video in adaptive streaming. Based on the possible bandwidth savings and resulting QoE variations, the algorithm decides on to adapt the bitrate dynamically and accordingly maximizes viewers' QoE. Since the proposed adaptation approach constructs an optimal path across the segments for adaptation, it also achieves QoE fairness among multiple clients in shared bandwidth environments. The experimental evaluation carried over real-world wireless network environments demonstrate that the proposed adaptation approach can maximize viewers' QoE in adaptive streaming, especially, under highly variable throughput conditions.