时间抖动约束的最优视频平滑

Optimal Video Smoothing under Jitters Constraints

描述了在时间抖动约束下视频平滑的最优算法．该算法假定视频流的码流大小为先验知识，在提前或推迟时间抖动约束下使得视频流的峰值和码率波动降低到最小．它实际上是在视频累积码流的约束通道中找到一条最短路径．通过实验模拟，探讨了该算法各视频流在不同抖动约束下的平滑率曲线，最优平滑方法与跳变窗口平滑方法加以对比．表明了该算法可以有效地降低视频流的峰值和码率波动，适用于对时延抖动敏感的实时视频传输．

The article describes a video optimal smoothing algorithm under jitters constraints, in which sizes of video bit streams are supposed as known in advance and peak and variance of video are made to be the smallest under ahead/delay jitters constraints. In reality, it is to find the shortest path with video accumulation constraints. Also, the article discusses properties of the algorithm in detail by simulations, such as: video smoothing curves with different jitters; comparison between optical smoothing and jump window smoothing; effects of different jitters constraints on peak and variance of videos, etc.