The Study And Algorithm Analysis Of The H.264 Video Compression Standard And The Optimization Research Of H.264 Encoder

Abstract In this paper, a discussion of the emerging video coding standard, H.264. is carried out. As a new video codec standard, H.264 applies many mostly used technologies, it provides fairly good coding quality while achieving a better coding efficiency and more simple syntax expression, and it is the most efficient video coding standard at present. Comparing to prior video codec standards. H.264 seems will he used in more application areas in the future.The H.264/AVC design covers a Video Coding Layer (VCL), which efficiently represents the video content, and a Network Abstraction Layer (NAL). which formats the VCL representation of the video and provides header information in a manner appropriate for conveyance by particular transport layers or storage media. Three are many new technical used to H.264 codec: Intra prediction, multi-picture motion-compensated prediction, Variable block-sized motion compensation with small block size, Quarter-sample-accurate motion compensation. Integer DCT transform, Loop-filter, CABAC for entropy encoding, etc, all these new technical and their algorithms will be discussed in this paper. The so called rate-distortion optimization selection method being applied in H.264 for selecting the best mode between many choices is also discussed in this paper.The way to speed up the video encoder is also introduced in this paper. The key to perform encoder optimization is to make sure which part of the encoder is themost lime-consumption part. The way to locate the time-consumption part is discussed first. After that, three ways to perform The optimization is introduced : first of all. the way to use MMX/SSE2 technology lo speed up the core computing part is introduced; secondly, discussion on how to reduce the unnecessary calls of some I'unclions is taking place; at last, a novel fast motion estimation algorithm based on NCDS algorithm is proposed, which is provided not only more efficient than many famous fast motion estimation algorithm, but also maintains good performance comparing to the basic FS algorithm. Fxperimcntal result shows that after being optimized, the new encoder achieves a 2.5-3 speed up over the original one.