Congestion Control For Real-time Audio And Video Streams

With the vigorous development of information and communication technology and streaming media technology,the related technologies involved in data stream transmission,such as audio and video coding and decoding,data stream transmission protocol,congestion control,etc.,are becoming more and more mature.People's demand for instant messaging has expanded from the early exchange of text and picture information to the integration of audio,video,text and image.In recent years,accompanied by 4G or even the upcoming 5G communication technology,people have further requirements for instant messaging in daily life,such as whether the video quality is high-definition,whether the audio is high-fidelity,and further hope to carry out higher quality audio and video instant messaging.There are many detection standards for real-time audio and video quality,such as synchronization of audio and video,frame loss of video,noise echo and so on.There are many technologies affecting these detection standards,and one of the key ones is congestion control technology in the process of audio and video streaming transmission.Based on the analysis and research of real-time communication technology of audio and video streams,aiming at the problem of poor transmission reliability and lack of congestion control mechanism of UDP protocol in the field of instant messaging,and combining with a series of theoretical research and experimental records,this paper designs a set of congestion control algorithm for real-time audio and video streams on Android platform.This algorithm is transplanted and improved on the basis of traditional Google congestion control algorithm.In the initial stage of instant messaging,an initial bit rate is set based on experimental data,and then the bit rate is increased exponentially every certain time interval until a threshold of start-up bit rate is reached.This approach greatly reduces the congestion caused by excessive increase or decrease in the initial rate of instant messaging.Aiming at both the receiving end and the sending end of real-time audio and video communication,this paper proposes that the receiving end computes the queuing delay of the current packet transmission network based on the arrival delay of the receiving end,and then calculates the adjusted bit rate of the receiving end.The sender judges the current congestion situation of the network based on the packet loss rate of the receiving end,and calculates an adjusted bit rate of the sending end.Combining the two rates gives the final adjustment rate.Emphasis is laid on the analysis of the change of bit rate according to network state when real-time video streams minimize queuing delay and other streams compete for bandwidth channels.Finally,the experimental results show that the improved congestion control algorithm for audio and video streams can ensure the stability of audio and video streams in the initial stage of the establishment of instant messaging,and then complete the rate adjustment in real-time audio and video communication according to the bandwidth condition of the network with high efficiency and quality.