Estimation of the category of notch frequency bins of the individual head-related transfer functions using the anthropometry of the listener’s pinnae

Abstract

An authentic approach to present three-dimensional acoustical sensation to everyone is to provide head-related transfer functions (HRTFs) that are exactly adapted to each listener. This approach is called the individualization of HRTFs. In order to generate individualized HRTFs, the present study proposes a modeling method of HRTFs that extracts the frequency bins in which spectral notches (N1 and N2) are included. The frequency bins for N1 and N2 were classified into two categories based on the just noticeable difference in notch frequency with regard to the vertical angle perception of a sound image. Then, discriminant analyses were carried out as objective variables of the category of each of the N1 and N2 frequency bins and as explanatory variables of ten anthropometric parameters of pinnae. The results show that the categories of N1 and N2 frequency bins for the front direction can be estimated with accuracies of 83.1% and 77.8%, respectively. For the rear direction, the accuracies of the category estimation of N1 and N2 were 74.3% and 77.1%, respectively.