The Research Of AVS-M Video Coding Algorithms And The Realization On DM642

Technical progress and increasing of demand promote the development of audio and video industry. Audio and video coding technology is the base of audio and video industry. Our country has good industrial foundation in this domain. However, because of the lack of mastery of the important techniques and standards, related corporations have long been enslaved to foreign corporations or organizations. AVS-M (mobile video) is the seventh part of AVS which is put forward by our country independently. It aims at next generation multimedia applications. And it has significance in multimedia communication, digital multimedia broadcast and portable consumer electronics etc.The real-time implementation of video coding requires very high computation power and memory bandwidth. General-purpose processor is not competent for it. On the other hand, hardware implementation is difficult to develop and lacks of scalability. So special-purpose DSP gains extensive attention.The purpose of this dissertation is to implement an efficient real-time AVS-M video encoder on DM642. First, we introduce the algorithms in AVS-M standard. Then by analyzing the tradeoff between performance and complexity of each module, we make some improvements and simplifications to the algorithms. For motion estimation, the computational load is greatly reduced without sacrificing the coding efficiency. And for mode decision, although there is some lose in coding efficiency, the complexity is greatly reduced. After that, according to the technical features of TMS320 DM642 DSP, we elaborately design the program flow and data flow of the encoder. We make use of DMA to transfer data to alleviate the burden of DSP core. Further more we utilize many code optimization techniques to take full advantages of the parallel computation units.Experiments validate that the encoder can realize real-time video coding at the speed of 25 CIF pictures per second. Moreover it possesses good visual quality and high compression ratio. The research of this dissertation is useful for the spread and application of AVS-M.