How binaural room impulse responses influence the externalization of speech

Florian Wendt, Robert Höldrich, Marton Marschall

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

11 Downloads (Pure)

Abstract

Binaural reproduction is becoming popular as people consume radio, TV, and music over headphones out and about. One of the challenges when listening with headphones is that the sound image appears inside the head. Binaural renderers are meant to overcome this by using binaural room impulse responses (BRIRs) to reproduce at the ears the exact sound waves that one would hear when listening to the real sound source. Nevertheless, binaural reproduction often does no better than a standard stereo recording at getting the sound image out of the head. This is because our individual pinna, head, and torso are different to the one simulated by the renderer. Research has shown that when the features of individual HRIRs are accurately simulated with headphones, listeners report externalized images, e.g. [1]. If a dummy head is used instead, then the image may be externalized, but it is usually diffuse or localized close to the head; especially for the synthesis of sources that are directly in the front of the listener [2]. Another failing in binaural reproduction is the acoustical divergence of the synthesized room and the actual listening room, yielding negative influence on the externalization. Research has shown that congruent room impulse responses (RIRs) in terms of amount of reverberation and direct-to-reverberant energy ratio between synthesized and listening room are needed to generate an externalized image, e.g. [3, 4]. Similar to generic HRIRs, a divergent room yields diffuse sound images which are perceived either close or even inside the listeners head. This study focuses on the relative contribution of individual HRIRs and the congruency of RIRs on externalization.
Based on an externalization model we sketch a listening experiment where listeners are asked to rate the externalization of speech in a simulated environment. After
this, we present the results and discuss them in the last section.
Original languageEnglish
Title of host publicationFortschritte der Akustik -DAGA 2019
PublisherDeutsche Gesellschaft für Akustik e.V.
Publication date2019
Pages627-630
Publication statusPublished - 2019
Event45. Deutsche Jahrestagung für Akustik - DAGA 2019 - Stadthalle Rostock, Rostock , Germany
Duration: 18 Mar 201921 Mar 2019
http://2019.daga-tagung.de/

Conference

Conference45. Deutsche Jahrestagung für Akustik - DAGA 2019
LocationStadthalle Rostock
CountryGermany
CityRostock
Period18/03/201921/03/2019
Internet address

Cite this

Wendt, F., Höldrich, R., & Marschall, M. (2019). How binaural room impulse responses influence the externalization of speech. In Fortschritte der Akustik -DAGA 2019 (pp. 627-630). Deutsche Gesellschaft für Akustik e.V..
Wendt, Florian ; Höldrich, Robert ; Marschall, Marton. / How binaural room impulse responses influence the externalization of speech. Fortschritte der Akustik -DAGA 2019. Deutsche Gesellschaft für Akustik e.V., 2019. pp. 627-630
@inproceedings{a0aa5b45d32c4e97b0572377f590e79b,
title = "How binaural room impulse responses influence the externalization of speech",
abstract = "Binaural reproduction is becoming popular as people consume radio, TV, and music over headphones out and about. One of the challenges when listening with headphones is that the sound image appears inside the head. Binaural renderers are meant to overcome this by using binaural room impulse responses (BRIRs) to reproduce at the ears the exact sound waves that one would hear when listening to the real sound source. Nevertheless, binaural reproduction often does no better than a standard stereo recording at getting the sound image out of the head. This is because our individual pinna, head, and torso are different to the one simulated by the renderer. Research has shown that when the features of individual HRIRs are accurately simulated with headphones, listeners report externalized images, e.g. [1]. If a dummy head is used instead, then the image may be externalized, but it is usually diffuse or localized close to the head; especially for the synthesis of sources that are directly in the front of the listener [2]. Another failing in binaural reproduction is the acoustical divergence of the synthesized room and the actual listening room, yielding negative influence on the externalization. Research has shown that congruent room impulse responses (RIRs) in terms of amount of reverberation and direct-to-reverberant energy ratio between synthesized and listening room are needed to generate an externalized image, e.g. [3, 4]. Similar to generic HRIRs, a divergent room yields diffuse sound images which are perceived either close or even inside the listeners head. This study focuses on the relative contribution of individual HRIRs and the congruency of RIRs on externalization.Based on an externalization model we sketch a listening experiment where listeners are asked to rate the externalization of speech in a simulated environment. Afterthis, we present the results and discuss them in the last section.",
author = "Florian Wendt and Robert H{\"o}ldrich and Marton Marschall",
year = "2019",
language = "English",
pages = "627--630",
booktitle = "Fortschritte der Akustik -DAGA 2019",
publisher = "Deutsche Gesellschaft f{\"u}r Akustik e.V.",

}

Wendt, F, Höldrich, R & Marschall, M 2019, How binaural room impulse responses influence the externalization of speech. in Fortschritte der Akustik -DAGA 2019. Deutsche Gesellschaft für Akustik e.V., pp. 627-630, 45. Deutsche Jahrestagung für Akustik - DAGA 2019, Rostock , Germany, 18/03/2019.

How binaural room impulse responses influence the externalization of speech. / Wendt, Florian; Höldrich, Robert; Marschall, Marton.

Fortschritte der Akustik -DAGA 2019. Deutsche Gesellschaft für Akustik e.V., 2019. p. 627-630.

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

TY - GEN

T1 - How binaural room impulse responses influence the externalization of speech

AU - Wendt, Florian

AU - Höldrich, Robert

AU - Marschall, Marton

PY - 2019

Y1 - 2019

N2 - Binaural reproduction is becoming popular as people consume radio, TV, and music over headphones out and about. One of the challenges when listening with headphones is that the sound image appears inside the head. Binaural renderers are meant to overcome this by using binaural room impulse responses (BRIRs) to reproduce at the ears the exact sound waves that one would hear when listening to the real sound source. Nevertheless, binaural reproduction often does no better than a standard stereo recording at getting the sound image out of the head. This is because our individual pinna, head, and torso are different to the one simulated by the renderer. Research has shown that when the features of individual HRIRs are accurately simulated with headphones, listeners report externalized images, e.g. [1]. If a dummy head is used instead, then the image may be externalized, but it is usually diffuse or localized close to the head; especially for the synthesis of sources that are directly in the front of the listener [2]. Another failing in binaural reproduction is the acoustical divergence of the synthesized room and the actual listening room, yielding negative influence on the externalization. Research has shown that congruent room impulse responses (RIRs) in terms of amount of reverberation and direct-to-reverberant energy ratio between synthesized and listening room are needed to generate an externalized image, e.g. [3, 4]. Similar to generic HRIRs, a divergent room yields diffuse sound images which are perceived either close or even inside the listeners head. This study focuses on the relative contribution of individual HRIRs and the congruency of RIRs on externalization.Based on an externalization model we sketch a listening experiment where listeners are asked to rate the externalization of speech in a simulated environment. Afterthis, we present the results and discuss them in the last section.

AB - Binaural reproduction is becoming popular as people consume radio, TV, and music over headphones out and about. One of the challenges when listening with headphones is that the sound image appears inside the head. Binaural renderers are meant to overcome this by using binaural room impulse responses (BRIRs) to reproduce at the ears the exact sound waves that one would hear when listening to the real sound source. Nevertheless, binaural reproduction often does no better than a standard stereo recording at getting the sound image out of the head. This is because our individual pinna, head, and torso are different to the one simulated by the renderer. Research has shown that when the features of individual HRIRs are accurately simulated with headphones, listeners report externalized images, e.g. [1]. If a dummy head is used instead, then the image may be externalized, but it is usually diffuse or localized close to the head; especially for the synthesis of sources that are directly in the front of the listener [2]. Another failing in binaural reproduction is the acoustical divergence of the synthesized room and the actual listening room, yielding negative influence on the externalization. Research has shown that congruent room impulse responses (RIRs) in terms of amount of reverberation and direct-to-reverberant energy ratio between synthesized and listening room are needed to generate an externalized image, e.g. [3, 4]. Similar to generic HRIRs, a divergent room yields diffuse sound images which are perceived either close or even inside the listeners head. This study focuses on the relative contribution of individual HRIRs and the congruency of RIRs on externalization.Based on an externalization model we sketch a listening experiment where listeners are asked to rate the externalization of speech in a simulated environment. Afterthis, we present the results and discuss them in the last section.

M3 - Article in proceedings

SP - 627

EP - 630

BT - Fortschritte der Akustik -DAGA 2019

PB - Deutsche Gesellschaft für Akustik e.V.

ER -

Wendt F, Höldrich R, Marschall M. How binaural room impulse responses influence the externalization of speech. In Fortschritte der Akustik -DAGA 2019. Deutsche Gesellschaft für Akustik e.V. 2019. p. 627-630