Subjective comparisons of auditory distortion

To assess the accuracy of comparing different levels of distortion when levels are well above threshold, judgments were elicited from 340 subjects using two test tones of similar harmonic distribution in terms of frequency. Even though the total level of distortion in one of the tones had been reduced, subjects could not invariably select the less distorted tone in a comparison. This was true even for those subjects who stated that they could tell that the two tones sounded different from each other. There are implications concerning the validity of subjects' reports in making auditory estimations of 'quality' between signals.