面向应用层QoS的实时视频优化码率控制

基于传统视频信源编码在应用层QoS的联合信源信道优化，研究了信源编码和信道编码之间合理分配带宽以及信道与信源联合编码下的有效码率控制问题．依据视频编码的实时可变帧率下的VBR视频网络传输信道模型，提出了网络接入带宽统计复用基础上的优化信道码率分配；根据信道传输失真，推导出基于图像帧的全局信源信道联合失真估算递归模型，按照模型运算率失真优化下的跳帧参数和帧内宏块刷新率，实现整体的图像编码质量优化控制．相对于普遍采用的基于宏块的局部优化控制策略，所提出的优化码率控制策略能获取更一致的图像质量和更高的性能增益．

The Optimized Rate Control Strategy Based on Application QoS for Real-Time Video

The joint source-channel optimization based on the classical source coding with application-layer QoS, focuses on two challenges:to optimally allocate bit-rate between source and channel coding and to effectively achieve joint source-channel coding gain. This paper fulfills available rate prediction on multiple video source statistical multiplexing by the codec's buffer control, deduces a global joint source-channel distortion regressive estimation model, fulfills an optimized picture quality control on the frame-skip parameter and the optimal intra-macroblock refreshment ratio. The proposed scheme might achieve more consistent picture quality and better performance gain, against the general optimal macroblock-based distortion strategy.