The main goal of this paper is to find the relationship between video content complexity and the objective quality, further to compare H.265/HEVC, H.264/AVC, and VP9 encoders in three modes based on complexity of the ...
详细信息
The main goal of this paper is to find the relationship between video content complexity and the objective quality, further to compare H.265/HEVC, H.264/AVC, and VP9 encoders in three modes based on complexity of the video content. There are three evaluation metrics: Peak Signal to Noise Ratio(PSNR), structural similarity(SSIM), and encoding time, which were introduced to compare the three video encoders comprehensively. Each video sequence has 250 frames, and their resolution is 1920x1080, differing in spatial and temporal complexity. An extensive range of video content complexity were selected in the video sequences. Objective quality evaluation based on the calculation of spatio-temporal activities is tested to validate its suitability as video content complexity indicator in quality assessment. According to the analysis of experiment results, all the encoded modes present an obvious same trend. On a fixed video resolution and bit rate, with the video complexity increasing, the objective quality is decreasing and the encoding times is increasing. It was also concluded that the objective quality of X265-slow is far better than other encoder modes of X264 and VP9. In terms of encoding time, the encoding time of H264 was shown to be less than both HEVC and VP9.
As telecommunication technologies developing,the multimedia service,especially video service,is more and more *** the confliction between high complexity of video coding and worse computation capacity of mobile termin...
详细信息
As telecommunication technologies developing,the multimedia service,especially video service,is more and more *** the confliction between high complexity of video coding and worse computation capacity of mobile terminals becomes a bottleneck in related research.H.264/AVC is one of the most popular international video coding *** reducing its computation is *** this paper, the distributions of the intra-prediction modes' costs are studied by experiments and compiling the statistics in both 16x16 and 4x4 blocks for luma in H.264/AVC I-frame *** the conclusion that the distributions of the cost can be modeled as F distribution is proven by Kolmogorov-Smirnov(KS) *** adaptive threshold prediction method is presented to predict the distribution's parameter of the *** the distribution's quantile is used to set up threshold in order to reduce the unnecessary *** to the experimental results in the last table,a good performance is shown with the method presented.
暂无评论