一种超低时延FEC实时视频编码算法  

A Ultra-Low Latency FEC Real-time Video Coding Algorithm

作　　者：杨梅 陈锋[2] 高廷金 俞萍萍 YANG Mei;CHEN Feng;GAO Tingjin;YU Pingping(School of Advanced Manufacturing,Fuzhou University,Fuzhou 362200,China;College of Physics and Information Engineering,Fuzhou University,Fuzhou 350108,China)

机构地区：[1]福州大学先进制造学院,福建晋江362200 [2]福州大学物理与信息工程学院,福建福州350108

出　　处：《电视技术》2023年第3期13-16,22,共5页Video Engineering

基　　金：实时视频多路并行传输中的码率控制与分流策略研究(61801120);2022年教育厅产学研专项(2022J01551)。

摘　　要：为满足5G实时视频传输的可靠性和低时延需求,提出一种面向低时延实时视频前向纠错(Forward Error Correction,FEC)传输控制算法。该算法根据实时视频帧到达情况和当前网络状况,建模FEC编码、缓冲区数据包与端到端时延之间的关系,优化当前时间段FEC编码器的最佳参数。实验结果证明,该算法能够使得98%的视频数据时延控制在100 ms之内,相较于静态前向纠错机制,时延平均减少了30 ms,且用户平均接收视频数据的有效损失率降低约2%。To meet the requirements of reliability and low latency for 5G real-time video transmission,this paper proposes a low latency real-time video Forward Error Correction(FEC) transmission control algorithm for low latency.The algorithm models the relationship between FEC coding,buffer packets and end-to-end delay based on the arrival of real-time video frames and current network conditions,and optimizes the optimal parameters of the FEC encoder for the current time period.The experimental results demonstrate that the algorithm can make 98% of the video data delay control within 100 ms,which is an average reduction of 30 ms compared to the static forward error correction mechanism delay,and the average effective loss rate of received video data for users is only reduced by about 2%.

关 键 词：视频传输 前向纠错 端到端时延 

分 类 号：TP311.1[自动化与计算机技术—计算机软件与理论]