节点文献

基于BSP-16处理器的H.264视频编码器的优化与实现

Optimization and Implementation of the H.264 Encoder Based on Equator’s BSP-16

【作者】 权立伟

【导师】 陈辉煌; 石江宏;

【作者基本信息】 厦门大学 , 电路与系统, 2007, 硕士

【摘要】 H.264/AVC是ITU-T和ISO/IEC联合推出的最新视频压缩标准,采用了近几年视频压缩方面的先进技术,以较高的编码效率和网络友好性成为新一代国际视频压缩标准。大量的试验表明H.264编码器的高效率是以更高的运算复杂度为代价的。实际应用中,恰当的使用H.264编码算法可以用低计算复杂度的算法来达到与高复杂度算法相近的编码效果。而在实际的视频编码器的开发中,需要对运算复杂度和编码效果之间进行兼顾考虑。本课题是通过参考当前PC平台上最成功的开源H.264编码器X264,在Equator公司以BSP-16多媒体DSP处理器为核心的Babelfish II多媒体开发平台上进行H.264的视频编码器的设计。为了能够降低编码器的运算复杂度,针对编码器中不同作用和实现方法的部分,需要采用不同的优化方式。对基本模块可以使用多媒体指令来优化,对于压缩算法可以通过研究快速算法来替代。本文基于以上思想进行优化,具体包括下述工作。由于X264中帧内最佳预测模式的选择使用的是全搜索的算法,这使得算法复杂度相当高。针对这种情况,本文提出了针对16×16亮度块和4×4亮度块的快速帧内预测模式选择算法。其中针对16×16亮度块的快速选择算法是基于宏块图像纹理属性而进行快速决策的,而针对4×4亮度块的快速算法是根据相邻预测方向代价相近的原理来进行快速选择的。为了更进一步提升编码器的性能以实现编码器的实时编码,本文采用BSP-16处理器的媒体指令集来对编码器进行优化。根据对编码器各模块的分析,使用FIRtree媒体指令集针对诸如运动搜索、运动补偿、整数变换、量化以及半象素插值等模块进行了优化,更进一步提高了编码器的执行速度。实验结果表明,通过对H.264编码器中运算密集模块进行算法和指令集的优化后,编码器的执行速度获得了大幅的提高。最终在Babelfish II开发平台上实现了CIF格式视频的实时编码。

【Abstract】 The newest standard approved by the ITU-T and ISO/IEC organizations is H.264/AVC, it represents a number of advances in video coding technology and becomes the newest international video coding standard because of both high coding efficiency and network friendliness. But the experiment results show that the improved coding efficiency comes with great complexity increase at the encoder. In fact, proper use of the H.264/AVC encode methods lead to roughly the same performances as the complex configuration. So both computational complexity and coding efficiency should be thought and made to the optimal trade-off in the implementation of actual coding system.In this paper, we present the design of H.264 encoder, based on the open source codec X264 ( one of the most successful H.264 encoders on the PC platform ) and BSP-16 development platform -- Babelfish II development board and iMMediaTools SDT. In order to reduce the complexity of the encoder, the fast algorithms are developed to replace the old one and the multimedia instruction set are used to optimize the basic modules. This dissertation is based on the above methods, and major works are as follows.Because in the X264, the full search algorithm is used to choose the best intra prediction mode. However it is very computationally expensive. To reduce the complexity of the intra prediction of the encoder, one fast intra-prediction mode selection algorithm for luma 16*16 block and one for luma 4*4 block are proposed. The fast algorithm for luma 16*16 block is using the macroblock properties to fast determine the best mode. And the algorithm for luma 4*4 block is using the modes with similar neighborhood direction have similar costs to fast determine.To further optimize the performance of the encoder, the multimedia intruction set of BSP-16 is used. Based on the analyse of the encoder, the relatively time-consuming modules such as motion estimate, motion compensation, integer transform, quantization and sub-pel interpolation are optimized through the FIRtree multimedia instruction set of BSP-16, which further improve the encoding rate.The experiment results show that after optimizing H.264 encoder at algorithm and instruction level, the computationally intensive modules in H.264 encoder was further optimized, and the encoder was significantly speeded up. And finally the optimized encoder can realize the real-time encoding (for CIF sequences) on the Babelfish II platform.

【关键词】 H.264编码器BSP-16
【Key words】 H.264EncoderBSP-16
  • 【网络出版投稿人】 厦门大学
  • 【网络出版年期】2008年 07期
  • 【分类号】TN762
  • 【被引频次】1
  • 【下载频次】48
节点文献中: 

本文链接的文献网络图示:

本文的引文网络