HDTV, UHDTV, 3D, HEVC, 4K… Those hot topics can be seen everywhere in video industry. In this year’s NAB show, there is even a 8K HEVC realtime codec! However, is it really the bigger resolution of the video, the better quality of experience (QoE)? How about the interpolated video?
Based on the current network infrastructure, fast-pace involvement of video resolution has put too much burden to the limited bandwidth. PSNR based compression is still dominate the market, but will be soon fade out for sure, because there are too much constraints for it to correlate well with QoE. Structural SIMilarity (SSIM)? already been widely adopted, but may not good enough, especially cannot handle single source scenario or transcoding. Where is the right tool for video perceptual quality measurement?