Perceptual weights in auditory level discrimination

Abstract

Perceptual weights in level discrimination (also called intensity discrimination) were determined for 3-, 7-, 15-, and 24-component tone complexes with flat spectral envelopes using a correlational paradigm. Each frequency component was randomly and independently perturbed in level oneach presentation. For the target interval, frequency-component levels were additionally increased by the level increment to be detected, deltaL [= 201og10((p + deltap)/p), where p is pressure]. Weights were calculated from the across-trial correlation between the level perturbations for each frequency component and the interval chosen by the listener. Two conditions were investigated: (1) deltaL was equal across frequency components, and (2) deltaL increased progressively across frequency components. For both conditions, data for four listeners usually showed the greatest weight for the highest frequency component. The two-to-four highest frequency components generally were most important for level discrimination. The effect of increasing deltaL progressively with frequency was small and inconsistent. Additional measurements showed that flanking noise maskers designed to mask spread of excitation caused only small and generally unsystematic changes to the weights. Overall, these results indicate that listeners combine information across a wide range of auditory channels to arrive at a decision for level discrimination, but the weighting of channels appears to be suboptimal.

The temporal weighting of loudness: effects of the level profile

“…Thus, a separate logistic regression model was fitted for each combination of participant, level profile, and mean level (μ L or μ S ). For each model, the weights w i were normalized such that the sum of their absolute values was unity (see Kortekaas, Buus, & Florentine, 2003), resulting in a set of relative temporal weights for each listener, level profile, and mean level.…”

Does a rhythmic context have an effect on perceptual weights in auditory intensity processing?

“…Participants were informed that because of the random sampling of the segment levels, feedback could be counterintuitive in some trials (cf. Kortekaas et al, 2003). The next trial followed the response after an intertrial interval of approximately 2 s.…”

“…The 14 weights estimated for each participant in each experimental condition were normalised such that the sum of the absolute values was unity (see Kortekaas et al, 2003). Note that the analysis is concerned with relative weights, for example, the question of whether the level of the first noise segment has a greater influence on the decision than the level of the fifth segment, regardless of the general strength of the association between the segment levels and the decision (cf.…”

“…If the participant assigns a large weight to a particular component-that is, if attention is directed to the component (Berg, 1990)-there will be a strong correlation between the component's random level perturbation and the participant's response. If on the other hand the component is unimportant for the decision, the responses will be statistically independent of the random variation (Kortekaas, Buus, & Florentine, 2003;Lutfi, 1995;Richards & Zhu, 1994). According to the attentional pulse hypothesis, temporal segments in an auditory stimulus to which attention is directed should have a larger influence on loudness perception than temporal segments presented at rhythmically unexpected points in time.…”

“…For example, Kortekaas et al (2003) found that in an intensity discrimination task involving multitonal complexes, participants placed a larger weight on the level information from the higher frequency components. In contrast, Green (1993) reviewed studies in which the accuracy for detecting an intensity increment on one component of a multitonal complex was measured and concluded that accuracy is highest for the central frequency component.…”