Enhancements to the opinion model for video-telephony applications

Abstract

In this paper, we show how the proposed model in ITU-T Recommendation G.1070 "Opinion model for video-telephony applications" cannot model properly the perceptual video quality, especially in the low bit rate range, due to the great variation of MOS values depending on video content. In this work, we present different enhancements to the model, allowing a much better approximation to the perceptual MOS values, knowing only the subjective movement content in the video application, classified in "Low", "Medium" or "High". Studies were made for more than 1500 processed video clips, coded in MPEG-2 and H.264/AVC, in bit rate ranges from 50 kb/s to 12 Mb/s, in SD, VGA, CIF and QCIF display formats. Video clips subjective quality was estimated using one of the quality metrics standardized in ITU-T Recommendation J.144 and ITU-R Recommendation BT.1683.