2009LiriasRequires access

Binaural Sound Source Localization in Real and Virtual Rooms

Monika Rychtáriková, Tim Van den Bogaert, Gerrit Vermeir, Jan Wouters

Open publisher page 30 citations

Abstract

On the basis of a statistical evaluation of the binaural sound source localization performance during listening tests by human subjects, it is shown that the convolution of a measured head-related transfer function (HRTF) with the room impulse response generated by a hybrid image source model with a stochastic scattering process using secondary sources provides an adequate model for the prediction of binaural room impulse responses (BRIR) for directional sound localization in the frontal horizontal plane. The source localization performance for sound stimuli presented to test subjects in a natural way was compared to that presented via headphones. Listening via headphones tends to decrease the localization performance only, and only slightly when localizing narrow-band high-frequency stimuli. Binaural sound presented to test subjects via headphones was generated on the basis of measurements and simulations. Two different headphone conditions were evaluated. The overall localization performance for simulated headphone sound obtained using a simulated BRIR was found to be equivalent to the one using a measured BRIR. The study also confirms expectations that the deterioration of sound source localization performance in reverberant rooms is closely linked to the direct-to-reverberant ratio for given loudspeaker and listener positions, rather than to the reverberation time of the room as such.

About this research paper

What this paper is about

On the basis of a statistical evaluation of the binaural sound source localization performance during listening tests by human subjects, it is shown that the convolution of a measured head-related transfer function (HRTF) with the room impulse response generated by a hybrid image source model with a stochastic scattering process using secondary sources provides an adequate model for the prediction of binaural room impulse responses (BRIR) for directional sound localization in the frontal horizontal plane. The source localization performance for sound stimuli presented to test subjects in a natural way was compared to that presented via headphones. Listening via headphones tends to decrease the localization performance only, and only slightly when localizing narrow-band high-frequency stimuli. Binaural sound presented to test subjects via headphones was generated on the basis of measurements and simulations. Two different headphone conditions were evaluated. The overall localization performance for simulated headphone sound obtained using a simulated BRIR was found to be equivalent to the one using a measured BRIR. The study also confirms expectations that the deterioration of sound source localization performance in reverberant rooms is closely linked to the direct-to-reverberant ratio for given loudspeaker and listener positions, rather than to the reverberation time of the room as such.

Why it matters

OpenAlex reports 30 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

On the basis of a statistical evaluation of the binaural sound source localization performance during listening tests by human subjects, it is shown that the convolution of a measured head-related transfer function (HRTF) with the room impulse response generated by a hybrid image source model with a stochastic scattering process using secondary sources provides an adequate model for the prediction of binaural room impulse responses (BRIR) for directional sound localization in the frontal horizontal plane. The source localization performance for sound stimuli presented to test subjects in a natural way was compared to that presented via headphones. Listening via headphones tends to decrease the localization performance only, and only slightly when localizing narrow-band high-frequency stimuli. Binaural sound presented to test subjects via headphones was generated on the basis of measurements and simulations. Two different headphone conditions were evaluated. The overall localization performance for simulated headphone sound obtained using a simulated BRIR was found to be equivalent to the one using a measured BRIR. The study also confirms expectations that the deterioration of sound source localization performance in reverberant rooms is closely linked to the direct-to-reverberant ratio for given loudspeaker and listener positions, rather than to the reverberation time of the room as such.

Key concepts: Binaural recording, Headphones, Head-related transfer function, Acoustics, Reverberation, Sound localization, Loudspeaker, Impulse response

Related papers

Back to paper searchBrowse research topicsOriginal source
Binaural Sound Source Localization in Real and Virtual Rooms — Research Paper | ScholarLens