PHYSICAL MODEL OF SOUND DIFFRACTION AND REFLECTIONS IN THE HUMAN CONCHA

An approximated physical model of the frequency transfer function of t he human concha is developed in this paper. This formulation includes diffraction reflection. and interference phenomena in the concha cavit y. The performance of the proposed diffraction/reflection model is com pared with that of the single-delay-and-add approximation by checking their predictions against the experimental transfer function of a meta l spiral-shaped diffracting/reflecting system. Results show that the d iffraction/reflection model performs considerably better at predicting both the absolute center frequency of spectral minima and the relativ e frequency spacing between them. The diffraction/reflection model is then applied to a realistic concha shape and its predictions an compar ed with experimental head-related transfer functions for azimuth- and elevation-varying sound sources. In this case, the model predicts the elevation-dependent spectral features related to the transverse dimens ions of the concha. Additionally, the diffraction/reflection model pre dicts that, because of sound diffraction, similar spectral features mu st be generated in the concha for sources at all azimuths within the f rontal part of the ipsilateral hemisphere. Experimental and theoretica l evidence supporting this prediction is presented. (C) 1996 Acoustica l Society of America.