Measuring and Predicting the Perceived Quality of Music and Speech Subjected to Combined Linear and Nonlinear Distortion

The results of experiments in which subjects rated the perceived quality of speech and music that had been subjected to various forms of both linear and nonlinear distortion are reported. Experiment 1 made use of artificial distortions (such as ripples in frequency response combined with peak clipping). Experiment 2 included both artificial distortions and real distortions introduced by transducers. The results were compared with the predictions of a new model based on a weighted sum of predictions for linear distortion alone and for nonlinear distortion alone. There was a very good correspondence between the obtained and predicted ratings. Correlations were greater than 0.85 for speech stimuli and 0.90 for music stimuli. It is concluded that the new model can predict accurately the perceived quality of speech and music subjected to combined linear and nonlinear distortion.