Objective assessment of audio quality using an auditory processing model

A new method for the objective assessment and prediction of the perceived audio quality is introduced. It represents an expansion of the speech quality measure qC , introduced by Hansen and Kollmeier (2000). It is based on a psychoacoustically validated, quantitative model of the ”effective” peripheral auditory processing by Dau et al. (1996a, 1997a). To evaluate the audio quality of a given distorted signal relative to a corresponding high quality reference signal, the model is employed to compute ”internal representations” of the signals, which are partly assimilated in order to account for cognitive aspects. The linear cross correlation coefficient of the assimilated internal representations represents the new audio quality measure PSM (Perceptual Similarity Measure). While PSM shows good correlations with subjective quality ratings if different types of audio signals are considered separately, a better accuracy of signal-independent quality prediction is achieved by another quality measure, PSMt, represented by the 5%-quantile of the sequence of instantaneous audio quality PSM(t). The new measures were evaluated using a large database of subjective listening tests that were originally carried out on behalf of the ITU1 and MPEG2 for the evaluation of various low bit-rate audio codecs. The results support the concept of amplitude modulation processing by a modulation filterbank and suggest a nonlinear relationship between the perceived instantaneous and overall audio quality. The observed good performance of PSMt in predicting subjective quality ratings is similar to the ITU standard BS.1387 (”PEAQ”) for most of the tested data. International Telecommunication Union Moving Pictures Experts Group