【作者】 杨飞；

【导师】 林家儒；

【作者基本信息】 北京邮电大学 ， 信号与信息处理， 2008， 硕士

【摘要】 数字视频多媒体技术由于其较高的质量和较强的抗误码性能,越来越受到人们的欢迎。但数字视频多媒体的数据量巨大,网络有限的带宽无法支持。需要首先对其进行压缩编码,然后才能进行存储和传输。因此,视频编码技术在数字视频多媒体技术领域起着至关重要的作用。AVS作为我国自主研发的音视频编解码标准,是目前最先进的音视频压缩编解码标准之一。它具有性能高、计算复杂度低、专利授权费用低等优点,有广阔的应用前景。DSP作为一种专用的数据处理芯片,以其开发周期短、使用灵活、代码可更新升级等特点,在视频编解码技术领域得到了广泛的应用。Blackfin系列DSP作为ADI公司与Intel公司联合开发的高性能的定点DSP产品,特别适合应用于对功耗、运算能力等方面要求比较高的音频、视频和通信领域。将AVS先进的编解码技术和稳定的DSP处理器相结合,达到高效的视频压缩性能,具有很高的工程意义和市场价值。本课题正是在这样的背景下提出的。本文首先对视频编码的基本理论和相关标准进行了介绍,然后重点研究了AVS视频编解码标准,包括AVS视频标准主要技术、AVS视频标准与其他主流视频编解码标准的对比、AVS视频标准的知识产权状况及AVS标准的发展前景等。同时对Blackfin系列DSP的特点、结构和软硬件开发环境等也做了简单介绍。本文在详细研究Blackfin系列DSP和AVS视频编解码标准特点的基础上,以ADSP-BF561评估开发板为核心完成编解码器硬件平台的设计。并在嵌入式ucLinux操作系统和Blackfin交叉编译环境下完成AVS视频解码算法在ADSP-BF561上的移植。本文以解码器优化为主线,针对ADSP-BF561平台的软硬件特点,对移植后的AVS视频编解码算法进行了一系列的优化,包括编译器优化、DMA优化、算法流程优化、存储结构优化和汇编指令优化等。由于核心解码模块的优化是解码器速度提高的关键,因此本文针对AVS解码器的特点,对AVS视频解码器中的子像素插值、整数(反)变换、去块效应滤波等核心功能模块进行了重点地分析和优化。测试结果表明,本文所做优化工作大大提高了解码器效率,且解码图像主观质量良好,达到了本课题预期的要求。更多还原

【Abstract】 The digital video multimedia technology receives increasing attention for its high quality and high error resilience. But it contains enormous data and couldn’t be supported by current network bandwith. Compression must be done to meet the requirements of storage and transport. So video coding technology is a key in the digital video multimedia technology field. AVS is a video/audio encoding and decoding standard developed by China independently, and it is one of the most advanced video compression standards that are available. It has many good advantages, including good performance, low complexity algorithm, low patent licensing fees and so on. We have reason to forecast that AVS can find many application fields.As one kind of special data processing chipes, DSP has been wildly used in video encoding/decoding technology field for its short exploitation period, flexibility and updatable features. Blackfin family DSP, which is developed by Analog Devices Inc. and Intel Corporation, is a fixed-point DSP product with high performance, and it is designed specifically to meet the upper computational demands and power constraint of audio, video and communications domain.It is significant in both engineering application and market value to combine advanced technique of AVS and stable DSP together which achieves a better video compression performance. This project is put forward in this background.The dissertation begins with the introduction of the basic theories and relative standards of video coding. And then it mainly researches AVS video coding standard, including main technology of AVS video standard, comparison between AVS video standard and other video coding standards, the IPR(Intellectual Property Rights) status and the development prospect of AVS video standard, etc. It also introduces the properties, structure and development environment of Blackfin family DSP. Employing ADSP-BF561 EZ-KIT LITE evaluation board and AVS video standard, the dissertation designs the CODEC(Coder-Decoder) hardware platform. Relying on embedded ucLinux OS(operating system) and Blackfin cross compiling environment, AVS video CEDEC algorithm is transplanted to ADSP-BF561 platform.According to the properties of ADSP-BF561, the dissertation optimizes the AVS video CODEC algorithm invoking compiler optimization, DMA optimization, algorithm flow optimization, storage structure optimization and assembly instructions optimization. The optimization is concentrated on decoder optimization. Because the optimization of core decoder module is a key to improve the speed of decoder. The dissertation mainly analyze and optimize the core modules of decoder, including sub-pixel interpolation, integer transform(inverse transform), deblocking loopfilter, etc. Test results show that the optimizing methods put forward in this dissertation significantly improve the efficiency of AVS decoder, and the subject quality of decoded image is good. It also meets the requirements of the design.更多还原