Mplementation And Optimization Of AVS+Encoder Based On OMAP3530

In recent years, increasing demands for video effects, and foreign monopoly videocoding standard core technology, and royalties expensive. China independentlydeveloped and reached international advanced audio and video coding standard AVS(Audio Video coding Standard) came into being. With the industrialization process ofAVS, AVS standard enhanced version of the standard AVS+launch and relatedindustries rapidly advancing, but AVS+reference software encoder relatively lowefficiency, and previous studies used in the H.264standard and can not be directlyapplied to the AVS+, for this paper studies AVS+encoder implementation andoptimization on the OMAP3530.This paper introduces in detail the core module performance comparison ofdomestic and international video coding standard, AVS+encoder and AVS+standardand H.264standard, analysis of encoder module coding efficiency, the efficiency ofintra prediction coding algorithm, presents a fast AVS+frame zero coefficient blocksMPPM and fast intra prediction algorithm based on coding first, zero block decisionalgorithm, the zero coefficient blocks can be directly encoded CBP syntax elements; fornon zero coefficient blocks, simplified method for the SAD value, through thecomparison of early exclude the possible intra prediction modes, the intra predictionmodes for rate distortion cost function value to compare, and then choose the best theintra prediction coding, test results show that the improved optimization algorithm canreduce the intra prediction coding complexity while improving the coding efficiency.After performing intra prediction algorithm optimization, design and buildOAMP3530research environment and modify it for different platforms differencesbetween different points, the success of the AVS+codec successfully transplanted intoOMAP platform. For post-transplant AVS+encoder at low OMAP3530platformoperating efficiency were optimized from C-level and assembly-level, C-leveloptimized separately from the data type adjustment, using inline functions, file-relatedoperations such as convenience, assembler level respectively encoder parameters,software pipelining, linear assembly, Cache optimized angles and optimized for low- efficiency modules, test results show that the coding efficiency of AVS+coderoptimized increased more than3.5times.