Individual and group loudness relations were obtained at a frequency in the region of impaired hearing for 100 people, 98 with bilateral cochlear impairment. Slope distributions were determined from absolute magnitude estimation (AME) and absolute magnitude production (AMP) of loudness; they were also derived from cross-modality matching (CMM) and AME of apparent length. With respect to both the means and the individual slope values, the two distributions closely agree. More than half of the measured deviations are less than 20%, with an overall average of -1.5%, meaning that transitivity is preserved for bilaterally impaired individuals. Moreover, over the stimulus range where cochlear impairment steepens the loudness function, both the group means and the individual slope values are clearly larger than in normal hearing. The results also show that, for groups of people with approximately similar losses, the standard deviation is a nearly constant proportion of the mean slope value giving a coefficient of variation of about 27% in normal and impaired hearing. This indicates, in accord with loudness matching, that the size of the slopes depends directly on the degree of hearing loss. The results disclose that loudness measurements obtained by magnitude scaling are able to reveal the operating characteristic of the ear for individuals.

The present article provides an analysis of loudness growth rates in normal and cochlear-impaired hearing for diverse groups with respect to age and backgrounds. Slopes are obtained from absolute magnitude estimation and magnitude production of loudness (measured values), and from cross-modality matching and absolute magnitude estimation of apparent length (predicted values). Consistent with an earlier study [R. P. Hellman and C. H. Meiselman, J. Acoust Soc. Am. 88, 2596-2606 (1990)], slopes calculated within the 15-30 dB stimulus range above the elevated threshold increase in size with the degree of hearing loss. The corresponding range of loudness values in normal hearing yields a slope near 0.60 independent of the threshold levels. This pattern of loudness growth holds for individuals and groups. But the intersubject variability of the slope is more labile, being larger across than within groups and larger for the measured slopes than for the predicted values. Determined from the predicted slopes, the coefficient of variation, sigma/m, is approximately constant in normal and impaired hearing ranging from 20%-27%. In contrast, sigma/m, obtained from the measured slopes, increases with the degree of hearing loss to a value of almost 50% for a 75-dB loss. The overall stability and systematics of the observed effects further demonstrate that the loudness-intensity relation can be specified with reasonable precision and accuracy from cross-modality matching.

An investigation of the relation among individual power-function exponents for 51 adults with normal hearing was undertaken. Three different psychophysical procedures were employed: absolute magnitude estimation (AME), absolute magnitude production (AMP), and cross-modality matching (CMM) between loudness and perceived length. From these procedures, loudness exponents obtained directly from measurements of AME and AMP of loudness were compared to exponents predicted from CMM and AME of perceived length. The means of the distributions of measured and predicted exponents were found to have an identical value of 0.56. Moreover, more than half of the differences between the predicted and measured exponents ranged from -.07 to +.09, giving measured deviations that extend from - 12.5 to 16%. The close agreement between the measured and predicted means, ranges, and distributions of exponent values implies that CMM combined with line-length information can be used with success to determine an individual's rate of loudness growth.

Loudness adaptation was measured for pure tones at 4, 12, 14, and 16 kHz. In three experiments, a total of 87 young listeners judged--by the method of successive magnitude estimation--the loudness of the tones over a 6-min exposure period. Thresholds were measured by an adaptive 2IFC procedure. Although earlier measurements had shown that adaptation near threshold increases with frequency, these new data reveal that the increase is especially marked at higher sensation levels. Thus, at 40 dB SL, over a 6-min period loudness declined by 18% at 4 kHz and by 94% at 16 kHz. Moreover, the 4-kHz tone remained audible for all listeners throughout the 6-min exposure period whereas the 16-kHz tone became inaudible for two-thirds of them by the end of the exposure period. Listeners with relatively low thresholds (< 50 dB SPL) at 16 kHz showed much less adaptation at 14 kHz than at 16 kHz and much less than listeners with relatively high thresholds (> 50 dB SPL) at 16 kHz; this latter group showed strong adaptation at 14 kHz. The marked loudness adaptation of steady tones at very high frequencies and relatively high sensation levels is associated with a restricted spread of excitation in the auditory system resulting from the steep rise of the threshold curve at the highest audible frequencies.

The onset-time difference ΔT needed to lateralize 30-msec dichotic tone bursts toward the leading ear was measured as a function of the frequency difference ΔF between the burst in one ear and the burst in the other ear. An adaptive procedure (a variation of BUDTIF) was used to concentrate judgments in the vicinity of 75% correct. Signal frequencies were centered at a geometric mean of 2000 or 6000 Hz. The level in the right ear was set near 25, 50, or 80 dB SPL. At each of the three reference SPLs, all tone bursts in the right ear were presented at the same loudness level. The level in the left ear was adjusted to center the image. Threshold ΔT remained approximately constant as ΔF increased up to the critical band, which is 300 Hz at a center frequency of 2000 and 1100 Hz at 6000 Hz. Beyond the critical band, threshold ΔT increased with ΔF. The critical bandwidth did not change with level for either listener, but the rise in threshold beyond the critical band became steeper at lower levels. These results provide further evidence that the dichotically measured critical band has the same width as the monaurally measured critical band, and that critical bandwidth is independent of sound pressure level. [Research supported by NIH Grant No. NS 07270.]