University of York
Improvements in the measurement and optimisation of head related transfer functions for binaural ambisonics
Abstract
dc:description.abstractIn recent years, the desire for spatial audio has surged with the inclusion of such technologies within popular streaming platforms and content creation workflows. Often presented over headphones as binaural audio, spatial audio allows a listener to experience a sense of externalisation and realism over and above traditional stereo playback. It is particularly suited to Virtual Reality; head mounted displays are fast becoming an affordable option to present 3 dimensional visual content and it is only logical that coherent accompanying audio should also exist. However, the challenge comes in achieving a life-like auditory image at minimal computational cost. Two things are needed to deliver high quality binaural audio: accurate measurement of the way in which humans interoperate a soundfield and a rendering engine capable of applying such methods to pre-prepared spatial auditory data. HRTFs (Head related Transfer Functions), are individual filters that describe the transfer function between a free-field source and the signals that arrive at a listener's ears. Ambisonics, a data storage and audio reproduction format based around the spherical harmonic functions, has become one of the leading approaches to such rendering engines. This thesis considers the capture and optimization of HRTFs for binaural-based Ambisonics. Spatio-temporal manipulations, a technique referred to as BiRADIAL, are shown to objectively improve the accuracy of binaural output through a perception-based spectral comparison model. A novel approach to HRTF measurement is then presented, capable of synthesising infinite far-field filters from just 50 real-world near-field measurements taken in under 7 seconds. Perceptual listening test results show an equivalence to the more traditional measurement approach despite the savings in time, cost and complexity.
Degree
thesis:*- Name dc:type.qualificationname
- Ph.D
- Level dc:type.qualificationlevel
- doctoral
- Grantor dc:publisher.institution
- University of York
- Year dc:date.issued
- 2019
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Armstrong, Cal
- Advisors dc:contributor.advisor
-
- Kearney, Gavin
- Damian, Murphy
- ORCID person:identifier.orcid
Rights
- Language dc:language.dcterms_iso639_2
- eng
Identifiers
dc:identifier.*- Identifier
- uk.bl.ethos.811414