EXPERIMENTAL DETERMINATION OF OPTIMAL PARAMETERS OF MOBILE TELECONFERENCES

Abstract

The article considers the problem of ensuring the availability and integrity of wireless subscribers in cellular and other wireless corporate networks. The article aims to determine the threshold values for the moment of failure of video transmission, quantitative parameters, artifacts, and the number of errors for the image. Show the dependence of the integrity of data transmitted in real-time on the characteristics of the environment. Two approaches were used to assess the quality of video information: qualitative (image recognition assessment) and quantitative (error measurement). Because the research program was written in the Kotlin programming language, a library written in Java or Kotlin was required. After searching the library, it turned out that only three libraries meet such parameters as reliability, relevance, and availability of documentation: Jaffree, Xuggler, and VLCJ. After gathering information, it was found that the most common screen extensions for desktops are 1366 × 768 and for phones—360 × 640. An error occurred that RTP did not support more than one connection. Also, the RTSP protocol could not pass the experiment on codecs other than MP4V. The experiment stopped earlier than necessary without error. Judging by the indicators, this was a very high CPU load. All other protocols were successfully tested, and results were obtained. During the experiments, we encountered various video anomalies. The worst was the video playback problem with the MJPG codec. Other anomalies were also identified: frame delay, incorrect frame rendering, white noise, and white noise mixed with frames. It is clear how up to 128 kbps experiments are successful and then begin to stop the video stream without error information. According to the results of experiments, the H.264 codec performs best.