基于延迟预测的自适应媒体播放算法

针对AMP-Live模型中存在的问题,提出一种基于报文延迟预测的自适应媒体播放算法（NEWAMP）,采用未来信道和缓冲状态的预测值作为视频报文播放速率调整的依据,将速率变化的程度进一步细化,同时考虑应用要求的最大端到端延迟,提高算法性能,与传统播放算法相比,NEWAMP在保证报文因下溢和上溢而丢弃的概率足够小的前提下,缓冲延迟减小了约50%,而与普通AMP-Live方法相比,NEWAMP不仅减小了报文因下溢和上溢而丢弃的概率,还将缓冲延迟减小了约40%。实验结果证明了该算法的有效性。

Adaptive Media Playout Algorithm Based on Delay Prediction

Aiming at the drawbacks in AMP-Live model, a NEW Adaptive Media Playout(NEWAMP) algorithm based on packet-delay prediction is proposed. The main modification is that, the predicted value of the future channel and the buffer state may be used as the basis of playout policies. Moreover, after thining the playout speed and considering the maximum end-to-end delay, the algorithm performance may be improved further. In contrast to the classical media playout, NEWAMP algorithm decreases the average buffer delay more than 50 percent with low enough packet loss ratio of overflow and downflow. To the traditional AMP-Live algorithm, not only the packet loss ratio of overflow and downflow is decreased, but also the average buffer delay is deceased almost 40 percent. Experimental results show this algorithm is effective.