Roving level tone in noise masking

Abstract

The results of tone in noise detection experiments have long formed an empirical basis for the notion of the “critical band” and have led to the development of the “energy detection” model. In this study, the detectability of tones in bands of random noise was measured for conditions in which the overall level of the noise was roved randomly from interval to interval of each experimental trial. The functions relating masked tone threshold to noise bandwidth for the roved conditions were identical to those obtained when no rove was employed at “supracritical” bandwidths. At “subcritical” bandwidths, the detection thresholds were higher for the roved conditions, but were still lower than would be predicted from arguments based purely on energy detection. These results suggest that the traditional critical band and energy detection models should be modified to account for discrimination based on spectral shape or waveshape. [Work supported by NIH and NMRDC.]