SMPTE Component-Coded Digital Video — Picture Quality Assessments

David Bennett

Large-scale picture quality assessments of digital video signals have been performed by the SMPTE as part of its efforts to help achieve internationally compatible digital television standards. Observer panels composed of a large number of television specialists examined over 500 still video images using internationally accepted procedures for subjective testing. The experiment was designed to reveal the effects on picture quality of three sampling rates (768, 864, and 912 samples per line) and three chroma ratios (1/1, 1/2, and 1/4) in a system based on Y, R − Y, and B - Y signal components. In addition, a sub-Nyquist luminance rate of 1/2 was examined. Picture quality was judged in three contexts - directly at the system output, in a simulated NTSC environment at the system output, and in a simulated picture processing environment at the system output using color matting and picture expansion as the principal processes. Results show chroma ratios to be more important than sampling rate for direct display and color matte systems. When interfaced with the NTSC system, quality distinctions between chroma ratios are obscured. Even the sub-Nyquist rate produces picture quality comparable to NTSC. Given the complexity of organizing experiments of this nature, it seems unlikely they will be repeated in the near future. The results reported here are therefore unique and should be valuable in planning future digital television systems.

Print ISSN
Published
1981-10
Content type
Original Research
DOI
10.5594/J01363