| Due to the huge volume of digital video data,the transmission and store of the digital video require very high transmission bandwidth and very huge capacity. These high requirements have become the biggest obstacles to popularize digital video communication applications. At the same time,these obstacles promote the application of the video compression coding. Based on different applications and different compression tools,many international video compression standards have been established by the international organizations.For different applications,the compression,store and transmission of video data use different formats and standards. After the digital video signals are compressed by different formats and standards,these compressed video streams need to be transcoded in many cases. The transcoding of compressed video stream is to transcode a compressed video stream from one format or standard to another. And the transcoding technology can be divided into two types:one is to transcode the video stream from one format to another,which is called homogeneous transcoding;the other type is heterogeneous transcoding,that is to transcode the compressed video stream from one compression standard to another.Aim at the homogeneous transcoding,the second chapter concentrates on the format transcoding of compressed video stream compliant to the same compression standard. The format transcoding includes the frame size conversion,frame rate or field rate conversion,and interlaced/progressive video conversion. In this chapter,the frame size trancoding with arbitrary rational number ratio of L/M is put forward. And according to this arbitrary size conversion,a motion-vector reusing algorithm is put forward too. Because the motion vector reusing method can reduce the complexity to a great extend,while the video quality difference between this method and the new complete motion estimation method is small,the motion-vector reusing algorithm is widely adopted by many transcoders. In addition,compressed domain transcoder can skip the DCT/IDCT transformation processes,so as to reduce the complexity greatly. So the chapter introduces the compressed domain transcoding technologies,including DCT-domain image down-conversion (in case of the simple conversion ratio),DCT-domain motion-compensation and inverse motion-compensation.For the heterogeneous transcoding,the third chapter concentrates on MPEG-2 to MPEG-4 compressed video stream transcoding. Because of many advantages,MPEG-4 will be dominant in the video transmission,storing and editing. While manycurrent video streams are compressed by MPEG-2 standard. So it is necessary to research and develop the efficient MPEG-2 to MPEG-4 compressed video stream trancoder. The chapter firstly analyzes some key issues of trancoding design. Based on comparison of these two standards,the chapter puts forward five trancoding architectures. Comparing these five architectures,each is a compromise between complexity and video quality. And the retrieval algorithms of mismatched macroblock for DCT-domain requantization trancoder are introduced in the chapter. Many simulation results testified that the motion-compensation only for P-pictures transcoding solution is a better compromise between the complexity and video quality.In order to optimize the MPEG encoder,the fourth chapter focuses on the fast algorithms for two bottlenecks of encoder,DCT and motion estimation. And MMX?and SSE technologies are also used to speed up the DCT transform and the SAD operation needed for motion estimation. The fast DCT algorithm not only can reduce the multiplication number,but also can combine the post-scaling and matrix transpose needed for fast DCT with the quantization and scan processes,so as to speed up the whole MPEG encoder. In addition,the table-lookup fast quantization algorithm can further speed up the encoding process. For motion estimation,the spatial-temporal correlation based fast motion estimation algorithm is introduced. This algorithm makes use of the spatial and temporal correl...
|
PDF Full Text Request