首页 | 本学科首页   官方微博 | 高级检索  
     

IPTV点播业务中减少特技播放交互延迟的方法
引用本文:林镜华,雷为民,李凌南,白松.IPTV点播业务中减少特技播放交互延迟的方法[J].计算机应用,2009,29(11):2897-2900.
作者姓名:林镜华  雷为民  李凌南  白松
作者单位:1. 中科院沈阳计算技术研究所2. 中国科学院沈阳计算技术研究所
摘    要:优化交互响应时间是提高IPTV用户体验质量(QoE)的重要措施。为了减少特技播放过程的交互延迟,提出对媒体随机访问点建立索引的方法,以便媒体服务器通过查找索引表即可实现随机访问点的快速定位和关键帧提取。通过对MPEG-2 TS流的系统层结构和H.264 NAL层的结构分析,探讨了封装H.264视频的TS格式媒体文件的两种关键帧提取方法。鉴于IPTV媒体服务器存储的媒体文件数量庞大,采用Hash的方法建立节目信息索引表,以提高节目信息查找效率。最后给出了在媒体预处理过程中为点播内容建立节目信息索引表和随机访问点索引表以减少IPTV特技播放交互延迟的具体实现方案。方案可行性在与华为合作开发的IMS-based IPTV原型系统得到了验证,测试结果符合三重播放业务QoE的交互响应时间要求。

关 键 词:用户体验质量  索引机制  快速切换  随机访问点  传输流  H.264协议
收稿时间:2009-05-02
修稿时间:2009-06-12

Optimizing trick-play responsiveness in IPTV application
LIN Jing-hua,LEI Wei-min,LI Ling-nan,BAI Song.Optimizing trick-play responsiveness in IPTV application[J].journal of Computer Applications,2009,29(11):2897-2900.
Authors:LIN Jing-hua  LEI Wei-min  LI Ling-nan  BAI Song
Abstract:Optimizing interactive response time is an important measure for improving quality of users' experience. To reduce end-points process delay, client caching delay and acquisition delay of Random Access Point (RAP), a novel method based on building RAP index table was proposed. Media RAP acquisition and key frame abstraction can be quickly done by searching index table. With the analysis of MPEG-2 system layer and H. 264 network abstract layer, two methods of key frame abstraction were given. Whereas there are so many media files stored in media server, program information index built by Hash method was discussed, then program information could be acquired quickly. Based on the index tables of RAP and program information built in pre-processing phrase, trick-play implementation was depicted in detail. Furthermore, the feasibility was verified in the IMS-based IPTV prototype developed by Huawei Technologies Co., Ltd and our laboratory.
Keywords:Quality of Experience (QoE)  index mechanism  quick switch  random access point  Transport Stream (TS)  H  264 protocol
本文献已被 万方数据 等数据库收录!
点击此处可从《计算机应用》浏览原始摘要信息
点击此处可从《计算机应用》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号