摘要
针对不可控网络下的视频应用,提出一种实时跟踪每个用户网络状况进行精细动态码率传输控制的算法,为用户在不稳定的带宽下提供持续清晰流畅的全局最优的视频体验.通过简化W5/3提升小波,在运动补偿时域滤波算法中引入层次化多参考帧预测,并采用基于优先级的传输层流打包方法以及基于块位长的嵌入式零块编码算法,形成适合精细动态码率传输和控制的完整小波视频编码方案.经过测试对比分析表明,该方案能够有效降低视频播放过程中的停顿、缓冲、中断及视频画面中的马赛克、拖影等严重影响用户体验的现象,改善用户视频体验效果.
For the video application in uncontrollable network,a new algorithm is proposed to track each user′s network bandwidth in real time and dynamically control transmission bit rate in fine granularity in order to provide a sustained,clear and fluent experience for video audiences with general optimization in the unstable network. W5/3 lift wavelet algorithm was simplified and the hierarchical multi-reference-frame prediction method was introducedinto MCTF(motion compensated temporal filtering); adopt the transmission layer stream packing method was adopted according to the data priority; the wavelet embedded block coding method was used based on bit-length search. All of these techniques together form the whole wavelet video coding scheme which is suitable for the dynamic bit rate transmission and control in fine granularity. After test and verify,it can reduce the phenomenon which seriously affecting the user experience such as pause,buffering,interrupt when playing and the mosaic or ghosting in the video frames. It can achieve higher user experience.
出处
《华中科技大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2010年第7期99-102,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(60704018)
关键词
小波视频编码
动态码率
多参考帧预测
媒体传输
媒体控制
实时跟踪
wavelet video coding dynamic bit rate multi reference frame prediction media transmission media control real-time tracking