Research On Digital Video Transcoding And Scalable Encoding Technique

With the rapid development of Internet technology, multimedia applications based on Internet, such as video on demand (VOD), network shopping, videoconference, remote education, field pickup and so on have emerged in endlessly. Video communication has become the most challenging thesis in these multimedia applications because of its continuous transmission, large quantity of data and requirement for real-time and reliability,Various network and receiving terminals require video stream should have the abiltiy to adapt itself to various transmission, decoding and display requirements. This paper introduces the traditional scalable encoding technology in MPEG-2 TM5 first. This technique encodes bitstream in"layer"format, nomally one video stream can have one"base"layer and one or more"enhance"layer. Encoding video data in layer format makes the decoder can chose a part of encoded stream to decode. Decoder can decode only the base layer data to get basic quality stream, and also can decode base layer and one or more enhance layer to get high quality video.Traditional scalable encoding technique can only decompose stream into limited layers, and it is lack of flexibility, while video transcoding can provide more flexible changement of bit rate for compressed video. Video transcoding is a process of converting a previously encoded video bit-stream into another bit-stream with a lower bit-rate, a different display format, or a different encoding mode, in this process complete decoding or reencoding is not needed. Video transcoding can be divided into homogeneous, heterogeneous, for error recovery and for multimedia communication design. This paper analyzed in detail the transcoder structures of bitrate, spatial definition and temporal definition downconversion in homogeneous video transcoding, and the production and compensation of drift error. In allusion to two most important factors evaluating transcoding performance: transcoding speed and quality of transcoded stream, three new arithmetics were put forward: one is the use and redetermination of macroblock's encoded mode which can accelerate the transcoding process; another arithmetic is refreshing intra macroblocks according to the motion vectors which can improve the quality of transcoded stream; the other method is refreshing intra macroblock according to not only motion vector but also residual information which can improve the quality as well as reduce the bit rate oftranscoded stream. This paper demonstrated the theory of these proposed arithmetics, gave out corresponding experimental results. And this paper also introduced hetergeneous, for error recovery and for multimedia communication design transcoding technologies.In order to conquer the limitation of traditional scalable encoding technology, the recently put forward arithmetic: FGS is an effective arithmetic to realize the finely scalarity of video quality. It is supposed to be used in network transmission situation, especially low bitrate video transmission field. Wavelet transform is another hot spot in signal processing, its multiresolution analysis property can help to realize spatial and temporal definition scalable video encoding. This paper took these two prospective arithmetics as advanced scalable encoding technology to do some theoretic analysis and algorithmic research.