Video Stream Processing Platform And Optimized

With the rapid development of video stream processing technologies, our life has been much improved. Because of the mass data quantum of video stream, there is a trade-off between improving the efficiency of compression & decompression and reducing the cost of systems. The main objective of this paper is to solve the problem of real-time processing of video stream's compression &decompression on MAP-CA DSP.In the past, DSP and multimedia processor is short in real-time handling of MPEG2 video's coding and decoding. The bottleneck of performance include: ①The ability of handling data;②The ability of transfering data, if the necessary data is not in local memory, the processor cant work; ③The dependency of data, parallel handling the data which be known beforehand. The MAP-CA DSP which is made by Equator Inc provide a hardware base of handling video stream: ①VLIW core CPU; ②DataStreamer DMA controller; ③Variable Length Encoder/Decoder. But now, MPEG2's compressing and decompressing algorithms and mechanism is inviable to MAPC-CA's architecture, the hardware can not efficiently work. Aim at these mostly questions of MPEG2's real-time handling, MPEG2's video coding and decoding system based on MAP-CA have been implemented. MPEG2's key algorithms have been improved. New methods have been presented. For example: quantization and inverse quantization, DCT and IDCT, motion compensation. The waiting time will be reduced if the data is transferred by DataStreamer DMA controller.The system test indicate the performance of MPEG2's coding and decoding systemhad been improve 484% after being optimized. The ability of video handling had beenimproved.