Video Compression Application Performance Optimization Study

Responding to multimedia technology rapidly developing and application field promptly enlarging, many multimedia compression and coding standards progress continuously. Among them,MPEG-4 plays an important role in the multimedia field. The MPEG-4 organization also worked together with ITU-T Experts Group on video coding, and developed the H.26L standard. One very important part of multimedia technology is video compressing technology. Now, a lot of excellent video compressing software is developing quickly. Among them, OpenDivx and TML9(which is the test model of H.26L standard) are two very important video compressing application program.The rapid development of multimedia technology and its very high requirement of data processing are challenging the computer technology. So, we must design multimedia application-oriented computer architecture to fit the data processing demand of video compressing programs, we analyzed the parallelism of two representative video compressing programs-OpenDivx and TML9,and drew a conclusion that it is effective to run video compressing application programs on the processor which uses parallel arithmetic units. This can instruct the design of multimedia processor.The present video compressing software is all based on MPEG-4 standard or its extended edition, but they have not implemented many advanced ideas of MPEG-4 standard. For example, the code model of MPEG-4 standard is object-oriented. This need us first segment video into several objects and then code the objects respectively. But the present video compressing software has not made segmentation on video. They simply partition the frame of a video into 8*8 blocks and then code the 8*8 blocks respectively. This hasn't achieved the performance that MPEG-4 standard expects. We analyzed the present video segmentation algorithms and deduced a video segmentation algorithm that fits the video compressing application. After introducing the video segmentation algorithm, the present video compressing program must be correspondingly modified in other sections. So, at last, we modified the motion estimation and compensation algorithm. Through large amount of experiments, we conclude that the video compressing program can achieve higher compression efficiency after using the improved algorithm.