MEASUREMENT OF SPEECH, AUDIO AND VIDEO QUALITY IN NETWORKS

Real-time video communication in the form of IPTV (Internet Protocol TV) and Internet TV over a WAN (Wide Area Network) is often subject to disruptive elements such as packet loss, excessive delay and delay variations called jitter. This research topic examines the quality of service impact that packet-loss has on video streams that use modern compression techniques. Specifically, the paper analyses the optimal Group of Picture size and grouping strategies of the three distinct frame types used by MPEG-4 encoded video files, in the face of simulated packet loss. The paper also seeks to further define the relationship between optimal Group Of Picture size and the motion and texture content of an encoded video file. Results are based on industry-standard objective quality metrics such as PSNR and SSIM. Technologies used include FFmpeg, Evalvid and the network simulator NS-2.Some secondary subjective tests are conducted using a streaming video session between two machines on a wired local area network running VLC Player with NetEm induced network impairments.