Implementation Of H.264 Decoder On DSP

Of late years, the developing of the telecommunication and the 3G protocol require higher performance to the video communication. And the video services and the low rates video communication apply more in the consuming and business marketing. The new service and the quality of wireless channels want the new video protocol having higher compression ability and robustness that types of video compression protocol are developing on both of them. In the middle of 2003, the ITU Recommendation H.264 have been accept by ISO and became a official international standard , the new standard bring new requirement, and it is urgent to carry out the new standard on to hardware .Along with the extremely developing of digital signal processing DSP, it is possible to implement video and audio conveniently, on the other hand ,the new video compression standard H.264 provide the proper means for communication . So it is significant to implement the H.264 CODEC on to DSP.Because the aim of JM model is to test the basic function of H.264 definition, the complexity of JM is too high to make a real product, even couldn't for realtime coding-decoding. So it is important to optimize the CODEC. The optimization can be divided into the coding optimization and the algorithmic optimization. This paper focuses on the optimization of coding. And the method is general to the SW optimization, and can improve the efficiency. The implement of these methods can apply on the optimization of H.264 CODEC, and the level of optimization can be seen to four levels: the structure level, the arithmetic level, the C code level and the assemble language level.And the aim of project is to develop a terminal decoder system fitted to DSP environment, the core codec of this system is H.264 , and the core DSP is TM1300 by Philips. The system can work by the control of embedded OS -PSOS. When the encoder and decoder work on the DSP successfully, we can use the PSOS to connect encoder and decoder together, and coordinate the two of them. The PSOS make encoder and decoder to two processes. The encoder and decoder process execute, wait, suspend separately, and the PSOS maintain the resource competition and share. So it is important to plan cooperation between processes of encoder and decoder, that make it can exchange data and work fluently. Then Adding the API and VideoIn/VideoOut, to make the selfloop successfully. In the end. we discuss the performance of CODEC and selfloop in detail.