Research And Implementation Of Rate Control Based On VVC

The rapid development of computer and mobile communication networks,the wide application of various video data collection terminals,the boom of emerging industries based on multimedia video data application,and the increasing demand of market and users for highquality video data have resulted in the production and consumption of a large amount of video data from various aspects,which has brought challenges to the coding,storage,and transmission of video.To cope with these challenges,on the evening of July 1,2020,the first version of the new generation of international video coding standard Versatile Video Coding was officially introduced at the 19th meeting of the Joint Video Experts Team(JVET).The rate control technology,as an indispensable part of the entire video coding process,coordinates the adaptation of encoder coding output with the channel transmission and storage.It is related to the utilization of channel resources and the final performance of video coding presented to clients.Therefore,it is still of great significance to study the rate control technology in the new generation of video coding.This paper studies the characteristics of screen content coding and improves the existing VVC rate control algorithm based on offline video coding scenarios and online video coding application scenarios,respectively.The main contents and contributions of this study are as follows:To solve the problems occurring in the offline screen content coding scenarios,such as the loss of control of encoder rate due to the heavy consumption of bits by mutant frame and a period of low-quality picture frame due to the large error of VVC original rate control algorithm in the bit allocation of the first frame,an improved rate control algorithm based on sliding window is presented.The algorithm first presents an improved partial sequence optimal bit allocation method for the first frame according to the influence of the video distortion of the first frame on the subsequent picture frame distortion.Then,the sliding window is adopted to realize the GOP and frame-level bit allocation according to the picture complexity estimation result.In addition,the picture complexity estimation result is also used to achieve key frame detection and divide the picture into key frame and non-key frame for bit control.Experiment shows that,compared with the 4.64%average rate control error of the VVC original rate control algorithm,the average rate control error of the improved algorithm proposed in this paper is reduced by 3.14%.In the application of online screen content-encoding,in addition to the limitation of low latency,it is also important to consider the adaptation of the encoder output rate to the channel bandwidth.To solve the adaptation to the channel bandwidth,a buffer zone of a certain size is usually introduced between the encoder and the channel.However,on the one hand,this will cause a certain latency.On the other hand,due to the characteristics of the screen content,the buffer zone is more likely to overflow and underflow.To solve these problems,this paper presents an improved low-latency rate control algorithm.The algorithm first improves the latency constraints of online video coding,takes the picture preprocessing time into consideration,and then,according to such constraints,adopts a less time-consuming picture complexity estimation method that is different from the offline application scenario.Meanwhile,to improve the accuracy of the rate control,the algorithm divides the picture into mutant frames,natural conversion frames,and static frames for processing separately according to the picture complexity.Experiment shows that,compared with the VVC original rate control algorithm with the buffer zone,the improved algorithm proposed in this paper improves adaptation of the PSNR-Y by 1.50 dB while reducing the rate control error by 5.49%.