Q. Huynh-Thu and M. Ghanbari (UK)
Video services, subjective assessment, perceptual quality.
Subjective assessment of broadcast television pictures has received extensive attention from experts in the field. For non-broadcast video, some standardized methodologies are in place but these are much less mature than the ones for television pictures. New applications such as mobile video have recently emerged and call for further study of appropriate subjective assessment methodologies. This paper compares two methods, the Absolute Category Rating (ACR) method and Double-Stimulus Continuous Quality Scale (DSCQS) method. The authors used the two methods for the quality assessment of low bit-rate, low frame rate and low-resolution video. We selected a variety of video content, codecs, bit-rates, frame rates and error profiles that are representative of mobile video applications. We show that the quality ratings given by viewers in the ACR (with hidden reference) method highly correlate with those ones given in the DSCQS method. ACR with hidden reference provided the same essential information than the DSCQS method for each test condition, whilst allowing the assessment of four times as many test files. This is a significant advantage as mobile video applications exhibit a wide range of degradations; thus an important number of test conditions are required to assess the performance of objective perceptual models designed to handle these applications.
Important Links:
Go Back
