Predicting perceived sharpness of broadband noise from multiple moments of the specific loudness distribution

Abstract

Equal sharpness contours for broadband noise were generated through matches made between a standard reference noise and comparison noise stimuli varying in spectral envelope. Based upon the results of the sharpness matching task, a set of 20 percussive broadband noise stimuli was constructed and presented to three subjects in a sharpness rating experiment using white noise as a standard stimulus. Predicting obtained sharpness ratings for these percussive broadband noise stimuli from first and second moments of the stimulus specific loudness distribution was more successful than conventional prediction based only upon the weighted first moment.