Research On Rate Distortion Optimization And Screen Content Coding Based On X264 In Conference Video Scene

Due to the outbreak of COVID-19,the increasing demand for telecommuting has driven the demand of the cloud video industry,and video conferencing has gradually been more widely used.H.264 has the advantages of high compression ratio and good network affinity.As an encoder conforming to the H.264 coding standard,x264 has good compression performance and is also an open source project,so its share is still very high.Although video conference has a great range of applications at present,x264 video encoder can not well meet the coding and compression requirements of video conference scenes.At present,there are some problems in conference video coding.Due to the rate distortion optimization only considering the information of the current coding unit,the impact of the current coding unit on subsequent coding units is ignored,At the same time,for the screen content sharing of conference video,because of its distinctive characteristics,that is,there are a large number of same video frames,but they are still transmitted.This is the optimization and improvement of x264 encoder in the following aspects for network conference video scene:(1)For the portrait part of the meeting video,because the portrait video is a scene with slow motion and strong correlation.Therefore,it is necessary to consider the influence of the current coding unit on the subsequent coding unit for rate distortion optimization.We estimate the influence of the current frame on the subsequent frame by pre reading the frame to adjust its coding parameters.If we obtain the subsequent frame,it will destroy the real-time performance of the conference video.In this thesis,we will use the characteristics of portrait video in the conference scene to estimate the importance of the current coding unit by establishing a backward time-domain propagation chain,compare to the forward time-domain propagation chain.The overall average performance decreases by 0.06% compared with forward search.(2)For the screen sharing part of conference video,most of the screen sharing contents are shared documents,slides and other contents,which are explained during sharing,resulting in the static state of the screen sharing contents for a long time.These frames do not need to be transmitted,so they can be selectively transmitted by analyzing the screen content,so as to greatly reduce the content to be transmitted,For 600 frames of completely static self collection sequences,the overall average improvement of coding performance is 40.58 %.The overall increase of normal screen sequence of self collection is 9.08%.At the same time,the screen content has many unnatural textures and edges.By closing the filtering module,the texture and edge of the screen content can be better preserved,so as to improve the quality of coding.The overall increase of normal screen sequence of self collection is 1.00%.