Search PubMed⌕ Search

Biomedical subjects

Zachary A Constan

Publications and source records attributed to Zachary A Constan.

2 recordsLinked to original sources

On the detection of dispersion in the head-related transfer function.

Because of dispersion in head-related transfer functions (HRTFs), the interaural time difference (ITD) varies with frequency. This physical effect ought to have consequences for the size or shape of the auditory image of broadband noise because different frequency regions of the noise have different ITDs. However, virtual reality experiments suggest that human listeners are insensitive to head-related dispersion. The experiments of this article test that suggestion by experiments that isolate dispersion from amplitude effects in the HRTF and attempt to optimize the opportunity for detecting it. Nevertheless, the experiments find that the only effect of dispersion is to shift the lateralization of the auditory image. This negative result is explained in terms of the cross-correlation function for head-dispersed noise. Although the broad-band cross-correlation function differs considerably from 1.0, the cross-correlation functions within bands characteristic of auditory filters do not. A detailed study of the lateralization shifts show that the experimental shifts can be successfully calculated as an average of stimulus ITDs as weighted by Raatgever's frequency-weighting function (Thesis, Delft, The Netherlands, 1980).

Adult↗

Interaural level differences and the level-meter model.

The interaural level difference (ILD) plays a significant role in sound localization. However, the definition of ILD for noise is open to some interpretation because it is not obvious how to deal with the inevitable level fluctuations. In this article, the ILD is interpreted as an energylike (time-integrated) measure of stimulus level, independent of other stimulus details-particularly interaural correlation. This concept is called the "level-meter model." The model was tested by measuring human ILD thresholds for noise stimuli that were interaurally correlated, or anticorrelated, or uncorrelated. An additional test (not involving lateralization) measured the threshold for level discrimination based on loudness. According to the level-meter model, all four thresholds should be the same. The experimental results showed that the predictions of the level-meter model held good to within about half a dB, although thresholds for level discrimination were systematically higher than ILDs. Among the ILDs themselves, thresholds were slightly higher for uncorrelated noise. The latter result could be explained by replacing the level-meter model with a loudness-meter model, incorporating temporal integration. The same model accounted for the bandwidth dependence of the threshold.

Adult↗