H.264解码端相关技术的研究及其在IP机顶盒中的应用

【作者】 肖巍；

【导师】 刘心松；

【作者基本信息】 电子科技大学 ， 计算机系统结构， 2006， 硕士

【摘要】 联合视频工作组JVT(Joint Video Team)于2003年正式提出了新一代视频压缩标准H.264/AVC。H.264/AVC标准支持从低带宽、高误码率的无线移动视频通信到高带宽、低误码率的有线视频广播等多种应用,因此,日益受到业界的关注。为了实现更高的编码效率,H.264/AVC标准采用了很多新的编码技术,如多参考帧预测、多尺寸编码块模式、1/4像素精度运动矢量、整数变换量化、基于内容的熵编码、新型帧内预测、去除方块效应的滤波器等。在解码还原的图像质量相同的情况下,H.264/AVC比MPEG-2和H.263的编码性能至少提高一倍以上。更高的视频压缩比,使得H.264/AVC编解码器的计算复杂度成几倍甚至十几倍的增长,需要消耗大量的时间和系统资源,同时,H.264针对网络流媒体的应用采用了分层结构、参数集等技术增加了网络适应性,提高H.264网络传输的效率和纠错能力。为实现H.264网络视频的实时播放,需要解决的问题包括:①H.264编码数据的传输、解析;②寻找高效的优化算法,减少解码器巨大的计算复杂度。本文针对项目中使用的基于x86架构的纯软件解码的IP机顶盒平台,从H.264网络媒体实时播放的角度,对支持H.264的流媒体协议栈和H.264解码器优化这两个方面进行了研究、应用。首先,本文针对H.264/AVC网络适应性的各方面技术进行了深入的分析,提出了一套支持H.264编码流传输、解析,具有可扩展框架的流媒体协议栈解决方案,其中着重介绍了H.264的分层结构、参数集、组帧模式等技术特点和实现方法,同时还介绍了流媒体协议栈具有可扩展功能的整体架构和各模块的接口设计。其次,本文讨论了基于x86架构IP机顶盒的H.264解码器的纯软件优化问题,在参考相关资料的前提下,从算法角度实现了一种运动补偿的优化算法,以及一种基于分组的可变长快速CAVLC熵解码算法。同时还从代码实现上给出了程序优化的一些方法,包括指令集的优化和程序结构上的优化等。对于优化后的解码器,重新设计了它的外部接口函数,并将此H.264解码器模块链接进机顶盒播放器的解码器管理模块中,从而达到扩展原有解码器功能的目的。之后我们对优化前后的H.264解码器进行了性能测试,结果显示,解码速度明显提高,实现了在目标平台上CIF图像的实时解码。最后,把协议栈和解码器整合进播放器,并用此播放器进行了功能测试和性更多还原

【Abstract】 H.264/AVC video coding standard is provided formally by JVT (Joint Video Team) in 2003, and it will be the second successful video compression standard after the MPEG-2/H.262 in future. The standard is designed to give a new technical solution for a broad range of applications, such as very low bit rate conversational services and entertainment quality broadcast, interactive video-on-demand services etc. For the enhancement of the coding efficiencies, the new standard adopts new tools as following: multiple reference pictures, variable block-size with seven block sizes in motion prediction, quarter-pixel accuracy for motion vector, short word-length integer transform, context-adaptive entropy coding and loop filter deblocking.H.264/AVC has two characteristic, "Network Friendliness" and high coding efficiencies, which make the complexity of the H.264/AVC codec several times higher than that of existing standards. Accordingly, the software-based real-time video codec demands faster-speed algorithms. At the same time, it needs more strategy to make use of the new specific "Network Friendliness" to enhance the efficiencies of the data transmission and errors correction ability. If want the software-based real-time video decoding become true, we should solve these problems, such as, implementing a new style of streaming networking stack and enhanceing the decoding efficient, the platform is a x86 structured IP Set Top Box(STB).Firstly, the dissertation introduced the content of the H.264/AVC’s“Network Friendliness”technology, and then implements a streaming stack to follow the standard. The stack has the ability to add new module to apply the latest standard for transmission new type media data. The implement of H.264 standard consists of the Parameter Set concept, NAL and packaging. All modules of the stack will be introduced in the dissertation.Secondly, a software-based real-time H.264/AVC is realized on x86 based IP STB by optimization for algorithms. Based on the previous research of others, this dissertation implements an optimization of algorithm for motion compensation. Context-adaptive variable length coding (CAVLC) is chosen as entropy coding, a high更多还原