Search for Zwicker’s specific loudness function.

Abstract

A key ingredient of Zwicker’s loudness model is the ‘‘specific loudness’’ function which relates loudness to excitation within a single critical band or Bark. For narrow-band stimuli near 1 kHz this relation is described by a power function with a mid-to-high level slope (exponent) about 20% lower than the overall slope of the loudness function. In normal hearing, a slope reduction can be obtained by partially masking a tone with an adjacent high-frequency band of noise. Implicit in these findings is the assumption that the noise effectively contains the high-frequency excitation evoked by the tone. Recent loudness data obtained by magnitude scaling procedures for 21 listeners with bilateral sloping high-frequency losses provide a validation of this assumption. Measured at a frequency beyond which earphone corrected thresholds increase by more than 50 dB/oct, loudness increases with level at a slower rate than at a lower frequency and in normal hearing. Shallower threshold functions, like high tone-to-noise ratios, have little, if any, effect on the tone’s rate of loudness growth. The overall shape and slope of the loudness functions can be predicted from Zwicker’s model of loudness summation modified to account for noise-induced hearing loss [M. Florentine and E. Zwicker, Hear. Res. 1 (1979)]. [Work supported by the VA Rehab. R&D Service and by NIDCD.]