A first instance of a reference video is stored. A primary video and a second instance of the reference video are simultaneously received. At least one quality of experience value that infers a perceptual quality of the primary video as received by a system is generated by comparing the first instance of the reference video to the second instance of the reference video on a pixel-by-pixel, frame-by-frame, basis and determining whether each pixel and each frame contained in the first instance of the reference video are contained in the second instance of the reference video.