Modeling spectro-temporal modulation perception in normal-hearing listeners

The ability of human listeners to detect and discriminate spectro-temporal ripples in a sound has been shown to be correlated with their speech intelligibility performance in several conditions. Thus, if a model would be able to account for the spectro-temporal processing limits in the auditory system, such a framework could be used to analyze the auditory processes contributing to and limiting speech intelligibility. Here, a model is presented that combines the concepts of the power spectrum model of masking (PSM; Patterson and Moore, 1986) with those of the speech based envelope power spectral model of masking (EPSM; Jørgensen and Dau, 2011). Effects of masking and changes in the signal-to-noise ratio in both domains are considered in the decision device of the model. The model was evaluated in experimental conditions of temporal, spectral and combined spectro-temporal modulation detection and discrimination using identical stimuli as input to the model as to the human listeners. The predictions were compared to the measured data obtained with 16 normal-hearing listeners. The model could account for the mean data in most of the considered conditions and might provide a valuable framework for investigating effects of hearing impairment both on spectro-temporal perception as well as speech intelligibility.