The quality of experience (QoE) of video streaming is degraded by playback interruptions, which can be mitigated by the playout buffers of end users. To analyze the impact of playout buffer dynamics on the QoE of wireless adaptive hypertext transfer protocol (HTTP) progressive video, we model the playout buffer as a G/D/1 queue with an arbitrary packet arrival rate and deterministic service time. Because all video packets within a block must be available in the playout buffer before that block is decoded, playback interruption can occur even when the playout buffer is non-empty. We analyze the queue length evolution of the playout buffer using diffusion approximation. Closed-form expressions for user-perceived video quality are derived in terms of the buffering delay, playback duration, and interruption probability for an infinite buffer size, the packet loss probability and re-buffering probability for a finite buffer size. Simulation results verify our theoretical analysis and reveal that the impact of playout buffer dynamics on QoE is content dependent, which can contribute to the design of QoE-driven wireless adaptive HTTP progressive video management.

中文翻译：

---

播放缓冲区动态对无线自适应 HTTP 渐进式视频 QoE 的影响

视频流的体验质量 (QoE) 会因播放中断而降低，而最终用户的播放缓冲区可以减轻这种情况。为了分析播放缓冲区动态对无线自适应超文本传输​​协议 (HTTP) 渐进式视频 QoE 的影响，我们将播放缓冲区建模为具有任意数据包到达率和确定性服务时间的 G/D/1 队列。由于块内的所有视频数据包在该块被解码之前必须在播放缓冲区中可用，因此即使播放缓冲区非空也可能发生播放中断。我们使用扩散近似分析播出缓冲区的队列长度演变。用户感知视频质量的闭式表达式是根据无限缓冲区大小的缓冲延迟、播放持续时间和中断概率导出的，有限缓冲区大小的丢包概率和重新缓冲概率。仿真结果验证了我们的理论分析，并揭示了播出缓冲区动态对 QoE 的影响是内容相关的，这有助于 QoE 驱动的无线自适应 HTTP 渐进式视频管理的设计。