Study And Implementation Of Parallel Algorithm In H.26L Video Encoder

Multimedia video coding is one of the most active fields in computer application, and it is determined to have better futures for the popularity of video entertainment. H.26L is a video coding standard proposed by ITU and ISO/IEC, it is the emerging standard for video communication and multicast in low bit rate conditions. H.26L has the advantages of high compression rate, good adaptation and etc. In order to gain good coding result, H.26L adopts many new algorithms, which increase the computing quantity greatly. In many practical applications of H.26L, real-time encoding is often required, but nowadays real-time can hardly be achieved on a single state-of-the-art computer.Processing video on high performance parallel computers has good potential and wide foregrounds. The studies in this field are getting more and more popular and some achievements were come out. It's of great significant to study the parallelism in H.26L encoder for its advantages. The study of this paper is based on the test model long-term 8.0(TML8.0) of H.26L. In this thesis, detailed parallel analysis and parallel performance testing for the encoder are given, and parallel implementation is included with simulation results. The main contents of this thesis are listed in the following:(1) Basic ways of video coding, the main features of H.26L vide encoder, primary theories and speed up performance evaluating methods of parallel design are introduced briefly, they are the foundations of this paper.(2) Based on the primary principle in parallel computing, many parallel implementation methods are brought forward considering the main features of H.26L. The parallel performances of these implementation methods are analyzed and tested according to the normal methods.(3) According to the analyses and test results above, two parallel methods are selected and implemented by using MPI. The simulation results on Dawning-2000 MPP machines of the two parallel methods are analyzed and compared.The analysis and simulation results show that parallel computing can greatly speed up the processing of H.26L encoder, and make it possible for the H.26L to be used in many video conditions.