Thulke, Mirko
(2015)
Head-related transfer functions calculated using the boundary element method for a head and torso simulator.
Master of Philosophy thesis, University of Liverpool.
![[img]](https://livrepository.liverpool.ac.uk/style/images/fileicons/text.png) |
Text
ThulkeMir_Oct2015_2032299.pdf - Unspecified Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (301MB) |
|
Text
ThulkeMir_Oct2015_2032299.pdf - Unspecified Access to this file is embargoed until Unspecified. Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (0B) |
Abstract
The work reported in this thesis focuses on the calculation of the head-related transfer functions (HRTF) using the boundary element method (BEM). The main aim is to investigate the potential in using boundary element methods to predict HRTFs for sound sources that are close to the human body. To assess the accuracy of BEM using SYSNOISE software, results from direct and indirect BEM have been compared to an analytical solution of a rigid sphere. These typically show close agreement between 100 Hz and at least 3 kHz. However, at higher frequencies, the non-uniqueness of the BEM solutions leads to irregular frequencies, which become a major source of error. To try and identify and reduce these errors, procedures were developed that could be used for BEM modelling of the head and torso simulator (HATS) for which no analytical model is available. The second part of the work focused on the creation of a three-dimensional model of a HATS for the BEM simulation. A laser scan of the HATS was imported into PRO-ENGINEER software but the complexity of the geometrical features of the ear required additional editing using BLENDER software. The final model was successfully meshed and imported into SYSNOISE. BEM simulations were used to calculate sound fields over the surface on the HATS between 500 Hz and 10 kHz for a point source that was 1 m in front of the HATS and two point source positions that were close to the side of the torso. This allowed visual confirmation of features such as the bright spot at the back of the head and pinna resonances. HRTFs from the BEM model were validated against measurements on the HATS in the anechoic chamber and, in general, close agreement was achieved between 500 Hz and 8 kHz. At 10 kHz, the discrepancy is attributed to irregular frequencies.
| Item Type: | Thesis (Master of Philosophy) |
|---|---|
| Additional Information: | Date: 2015-10 (completed) |
| Subjects: | ?? NA ?? |
| Depositing User: | Symplectic Admin |
| Date Deposited: | 12 Jan 2016 15:19 |
| Last Modified: | 17 Dec 2022 01:44 |
| DOI: | 10.17638/02032299 |
| Supervisors: |
|
| URI: | https://livrepository.liverpool.ac.uk/id/eprint/2032299 |
Dimensions
Dimensions
