Rate Control Algorithm For H.264and The Realization

Video compression has become an essential component of communication, computer application, broadcast and entertainment media, which has brought lots of visual enjoyment to the lives of the people. In this paper, we first analyzed different rate control algorithms which could be used in H.264. And then realized and optimized the H.264coding algorithm on DM642.First of all, many different rate control algorithms were deeply analyzed including the classic ones used in MPEG-2, H.263and MPEG-4. Then the rate control algorithm proposals for H.264were compared and analyzed, which are JVT-D030, JVT-F086, JVT-G012and JVT-O016. We also studied and compared many other algorithms for H.264. And in the end, we found that the rate control algorithms using appropriate frame skipping could obtain the best result especially in real time applications according to all the analysis.Second of all, according to the importance of frame interpolation in rate control using frame skipping, a new effective frame interpolation algorithm is proposed. In our scheme, global motion vectors (MVs) were used to enhance the accurateness of motion estimation in the frame interpolation using bi-directional motion estimation. And in the mean time, we corrected the outliers in the MVs, using the Box-and-Whisker method in accordance with the statistical result. The experimental results show that our scheme has achieved a better visual quality with higher PSNR in the interpolated frames.Finally, the H.264coding algorithm was realized and optimized on DM642using four ways:1. Three parallel tasks, collection, coding and transmission were realized by using DSP/BIOS. And in this way, the system efficiency was highly increased.2. The whole algorithm was simplified according to the algorithm complexity analysis. We also use a fast mode decision algorithm to replace the original one for the need of the real time application.3. The system inner and outer memories were assigned reasonably to different models in H.264in accordance with their usage.4. The code was revised and optimized using C/C++compiler and assembly optimizer. After these four ways, the whole system efficiency was highly increased.