Back to list

Numerical calculation of individual head-related transfer functions of human listeners

Authors Ziegelwanger, H.
Year 2017
Thesis Type Doctoral thesis
Topic Audio Signal Processing
Abstract Head-related transfer functions (HRTFs) describe the directional filtering of the incoming sound at the ear canals. Spectral and temporal features of HRTFs are determined by the individual geometric details of a listener’s head, torso and pinnae. Thus, listener-specific HRTFs are absolutely essential for an accurate sound localization in binaural 3-D audio reproduction systems. In contrast to the common acoustical measurements, where small microphones are placed at the entrances of the ear canals, HRTFs can be calculated numerically based on a listener‘s discretized geometry, captured by geometry acquisition techniques like laser-scanning or photogrammetric reconstruction. While acoustically measured HRTFs usually provide a sound-localization performance similar to that obtained in free-field listening, the performance obtained with numerically calculated HRTFs, however, depends on the quality of the geometric and acoustic model of the listener used for the numerical calculation. Apart from the problem of unclear requirements on the geometry, the computational effort of the numerical HRTF calculation is large and the calculation process lasts tens of hours. The aim of this Phd project is to address open issues in the calculation of HRTFs under psychoacoustic quality criteria and to find methods to decrease the computational effort in the numerical HRTF calculation process. The modeling of the virtual microphone, the discretization of the boundary surface (head and ear) geometry and existing geometryprocessing algorithms to reduce the computational effort in the BEM were investigated. Results were evaluated with a sagittal-plane soundlocalization model, a time-of-arrival model and in sound-localization experiments. To stimulate further research in the field of binaural audio, the BEM code and the time-of-arrival model were published open-source as Mesh2HRTF and in the auditory modeling toolbox.
Supervisors Höldrich, R.