Research On H.266/VVC Complexity Analysis And Interframe Algorithm Optimization

The H.264 / AVC and H.265 / HEVC video codec standards have achieved great success since their release,in order to improve the codec effect of higher resolution and higher bit depth video,a new generation of codec standard VVC is under the development.Based on the previous generation of H.265 / HEVC,VVC has added some new technologies to save half of the bit rate when encoding images of the same quality.This paper studies the new generation video codec standard VVC,focuses on the encoding performance and complexity of the new standard,analyzes and optimizes the quadtree plus multi-type tree partition,which accounts for the higher encoding complexity.In the process of formulating each generation of standards,it is necessary to test performance.The test results can more comprehensively show the performance improvement of the new standard relative to the original standard after the addition of new technology,and provide optimization algorithm research for standard setting.The VVC standard is in the research and development.There is no authoritative performance test literature for the VVC standard.The official performance test tool only provides a bit rate performance test algorithm.The test results are scientific,but it can still be further improved.This paper designs an algorithm combining compression bit rate and image distortion for performance testing.The algorithm performs a comprehensive calculation of the video encoding compression rate and the integrity of the encoded image to obtain a more scientific and effective performance test result.Experimental results show that VVC improves the encoding performance by about 20% compared to the previous generation standard HEVC,increase 0.7d B about image quality,but the encoding complexity increases by nearly 5 times.The performance test results found that the VVC standard has a large increase in terms of encoding complexity compared to the H.265 / HEVC standard.Therefore,it is necessary to analyze the complexity of the VVC encoder.Based on this,a complexity analysis algorithm is designed in this paper.The algorithm calculates the time consumed by the main functions in the encoding process and classifies the main functions,so as to find the module with the highest encoding complexity.Experimental results show that inter-frame prediction encoding accounts for 70% of the entire coding process,so it is necessary to optimize its complexity.Quadtree plus multi-type tree partition is a new technology in VVC.It adds multi-type tree partition based on the quadtree partition of HEVC,which makes block partition more flexible and improves encoding performance.It also greatly increases the encoding complexity.In this paper,by analyzing the unit division level in the VVC standard inter prediction process,this paper proposes an optimization algorithm for the division level limitation,and reduces the encoding complexity by ending the division early.The experimental results show that compared with the original algorithm,the optimized algorithm loses 1.35% of its performance while reducing the encoding complexity by 44%,the new algorithm greatly increases the practical value of the new standard.