Visually adaptive virtual acoustic imaging

Abstract

Virtual acoustic imaging systems give listeners the perception of sound images at locations where no sound sources exist. They achieve this with filters whose design incorporates the paths’ transfer functions from the sound sources to the listener’s ears. The success of a virtual acoustic imaging system depends greatly on how close the listener is the intended listener location. The small area of acceptable listener locations is a major limitation of virtual acoustic imaging systems.  Adaptable virtual acoustic imaging systems greatly enhance the usefulness of the system and a listener’s enjoyment by modifying the intended listener location by adapting the virtual acoustic imaging filters as the listener location changes. The work in this thesis looks at the performance of an adaptable virtual acoustic imaging system that utilises a video head tracking procedure to track the listener’s movements.  Computer simulations and subjective evaluations consider the size of the ‘sweet spot’ at a variety of listener locations.  Dynamic subjective tests also consider the update rate of the filters with and without incorporation of a simple video head-tracking algorithm. The computer simulations show that due to the change in geometry, listener head locations further from the inter-source axis result in a higher ‘sweet spot’ frequency range. At asymmetric arrangements, head shadowing decreases the robustness of the system at the contra-lateral ear and increases the robustness at the ipsi-lateral ear. The most important parameter affecting the ‘sweet spot’ is how closely the virtual acoustic image location corresponds to the system’s loudspeakers. Listener movements of about every 3 cm or less seem to require a filter update for image stability. The image processing computation time significantly affects the perceived stability of a virtual acoustic image.

More information

Identifiers

Catalogue record

Export record

ASCII CitationAtomBibTeXData Cite XMLDublin CoreDublin CoreEP3 XMLEndNoteHTML CitationHTML CitationHTML ListJSONMETSMODSMPEG-21 DIDLOpenURL ContextObjectOpenURL ContextObject in SpanRDF+N-TriplesRDF+N3RDF+XMLRIOXX2 XMLReferReference ManagerSimple Metadata