Binaural reproduction using multiple dipoles

  • Lacouture Parodi, Yesenia (Project Participant)
  • Møller, Henrik (Project Participant)
  • Rubak, Per (Project Participant)

Project Details

Description

With binaural technology it is possible to simulate a virtual environment where the listener perceives an acoustic source located at a position where no physical source exists.  Based on the assumption that our sound perception is controlled by the sound pressures at the ears, one can simulate any virtual source if complete control over the sound pressures at the left and right ears is achieved.  Normally, headphones are used to reproduce the binaural signals due to the perfect channel separation that they provide.  However, reproduction through headphones might not be practical or comfortable for some applications and therefore loudspeakers are preferred.  When reproducing binaural signals through loudspeakers, the signals that are to be heard only by the right ear are also heard by the left ear and vice versa.  This is called crosstalk.  It is possible to reduce -even eliminate - the crosstalk by adding proper filters to the reproduction chain.  These techniques are known as crosstalk cancellation. In general, crosstalk cancellation lacks robustness regarding several parameters in the physical setup, such as reflections, listener's position and head movements.  Using small angles between loudspeakers - generally referred as stereo-dipoles - increases the robustness of the binaural reproduction system, but it is also known for poor performance at low frequencies.  The project is investigating alternative configurations for binaural reproduction using loudspeakers, which can potentially overcome some of the aforementioned drawbacks. At the moment, a thorough examination of the possibilities and limitations of the stereo-dipoles is being carried out.
StatusFinished
Effective start/end date19/05/201031/12/2017

Funding

  • <ingen navn>

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.