Comparison of H.264/AVC, H.264 with AIF, and AVS based on different video quality metrics

Abstract

Video communication has a very important role nowadays. Also very efficient video coding techniques are developed in order to provide high quality video streams using all the available bandwidth. In this work we perform a comparison between three different video codecs: the ITU-T H.264, an improved ITU-T H.264 version including the AIF (Adaptive Interpolation Filter), and the AVS (Audio and Video Coding Standard) that is the Chinese standard video codec. The realized tests involve eight standard video sequences that are coded end decoded using the mentioned standards. Then, for each sequence we evaluate the video quality using four objective metrics: PSNR (Peak Signal-to-Noise Ratio), VQM (Video Quality Metric), SSIM (Structural SIMilarity), and the HSM (HVS -Human Visual System -Similarity Measure). While PSNR and VQM take into account the mathematical features of the video sequences, SSIM and HSM are based on the HVS theory. So it is possible to evaluate also the human eye perceived quality. The realized analysis is related to the relationship between the bitrate and the quality indexes. The conclusions of our analysis can give an evaluation about several codec performances and can be useful for developers to plan and manage their network. Also, taking into account the HSM scores, the AVS has better performance than the ITU-T video codecs.

Telecommunications in the ICT Age: From Research to Applications

“…Coding and encryption of video signal, to be used in video-communication systems, have been considered in the framework of a source coding activity. With the task to test the effects of the features implementation introduced by the video coding standard H.264, an extensive performance evaluation of spatial, temporal, and hybrid error concealment techniques was provided [58]. The quality of the recovered image was measured, and the comparison among the various schemes is developed, using the JM7.3 Reference Software, either in subjective or in objective terms [40].…”