Perceptual weights for loudness judgments of six-tone complexes

Abstract

Subjects with normal hearing (NH) and with sensorineural hearing loss (SNHL) judged the overall loudness of six-tone complexes comprised of octave frequencies from 0.25 to 8 kHz. The level of each tone was selected from a normal distribution with a standard deviation of 5 dB, and subjects judged which of two complexes was louder. Overall level varied across conditions. In the "loudness" task, there was no difference in mean level across the two stimuli. In the "sample discrimination" task, the two complexes differed by an average of 5 dB. For both tasks, perceptual weights were derived by correlating the differences in level between matched-frequency tones in the complexes and the loudness decision on each trial. Weights obtained in the two tasks showed similar shifts from low to high frequency components with increasing overall level. Simulation of these experiments using a model of loudness perception [Moore and Glasberg (2004), Hear Res. 188, 70-88] yielded predicted weights for these stimuli that were highly correlated with predicted specific loudness, but not with the observed weights.

Effects of relative and absolute frequency in the spectral weighting of loudness

“…Recently, Oberfeld et al (2012) obtained perceptual weights for loudness of noise made up of multiple bands in a condition where levels of the noise bands were equated to produce equal loudness. In line with the results of Jesteadt et al (2014), their listeners put highest weight on the lowest component of the complex. Oberfeld et al (2012) demonstrated that the weights could not be accounted for by assuming different slopes of the loudness function at different frequencies.…”

“…5.5] suggests that the model may underestimate the steepness of the loudness function at low frequencies. Jesteadt et al (2014) showed low frequency emphasis at low overall levels in both equal-SL and equal-SPL conditions. The lowest frequency component was presented at levels closer to quiet threshold than other components in the equal-SPL condition and at a higher SPL in the equal-SL condition.…”

“…With increasing level, however, the low frequency emphasis was reduced and high frequencies received more weight. Although the loudness model (Moore and Glasberg, 2004) failed to predict the observed weighting function or any change in relative weights with increasing level, Jesteadt et al (2014) showed that weights from the model correlated with the specific loudness predictions of the model, suggesting that the underlying specific loudness of the components of a broadband sound could be assessed behaviorally using the perceptual weights. An increase in high frequency weight with increasing level has also been reported for sample discrimination of intensity (Kortekaas et al, 2003;Leibold et al, 2009) and discrimination of spectral shapes (Lentz, 2006).…”

“…To further investigate summation of loudness across the auditory filters, Jesteadt et al (2014) used a six-tone complex, components of which were widely separated over a five-octave range (octave spaced frequencies, from 250 to 8000 Hz). Such a stimulus avoids the possible confounding effect of one component masking another and reducing its contribution to the overall loudness.…”

Physiology, Psychoacoustics and Cognition in Normal and Impaired Hearing

“…Berg 1990;Lutfi 1995;Richards 2002). Studies using this paradigm show listeners to have highly replicable, individualistic patterns of decision weights on frequencies affecting their ability to hear out specific targets in noise-what has been referred to as individual listening styles (Doherty and Lutfi 1996;Lutfi and Liu 2007;Jesteadt et al 2014;Alexander and Lutfi 2004). Unfortunately this paradigm is extremely time-consuming, rendering it ineffective for clinical use.…”