Effect of Spectral Shaping and Content on Loudness Discomfort

Abstract

The purpose of this investigation was to study the impact of spectral shape and content on thresholds of discomfort (TD) for listeners with normal hearing and listeners with hearing loss. Secondary to that purpose was to quantify binaural summation at high intensities across complex stimulus conditions for both groups of listeners. Forty subjects (20 with normal hearing, 20 with hearing loss) participated. Complex acoustic stimuli (multitone and continuous discourse) were filtered to have four spectral shapes: (1) flat spectrum, (2) long-term average speech spectrum, (3) reverse long-term average speech spectrum, and (4) theTD contour derived for each subject from pure-tone TD obtained with eight pure tones from 250 to 4000 Hz. The results suggest that (1) TD for complex stimuli are lower for subjects with hearing loss compared with those with normal hearing, suggesting increased loudness summation with this population; (2) binaural summation of approximately 6 dB (independent of stimulus type, filter shape, or spectral content), indicating that a correction of similar magnitude for bilateral hearing aid fittings is appropriate; and (3) TD obtained at 750, 1500, and 3000 Hz accounted for approximately 60 percent of the variance in the complex TD measures, suggesting that TD at these frequencies be used to set the output obtained from a hearing aid with a 90–dB pure-tone sweep as the input stimulus. Abbreviations: ANOVA = analysis of variance, FIR = finite-impulse response, FS = flat spectrum, LTASS = long-term average speech spectrum, OSPL90 = output obtained from a hearing aid with a 90–dB pure-tone sweep as the input stimulus, R-LTASS = reverse long-term average speech spectrum, TD = threshold(s) of discomfort, TD contour = spectrum derived from TDs obtained with eight pure tones from 250 to 4000 Hz