On detectable and meaningful speech-intelligibility benefits

The most important parameter that affects the ability to hear and understand speech in the presence of background noise is the signal-to-noise ratio (SNR). Despite decades of research in speech intelligibility, it is not currently known how much improvement in SNR is needed to provide a meaningful b...

Full description

Bibliographic Details
Main Authors: Whitmer, William M., McShefferty, David, Akeroyd, Michael A.
Format: Article
Published: Springer 2016
Subjects:
Online Access:https://eprints.nottingham.ac.uk/38746/
_version_ 1848795681272102912
author Whitmer, William M.
McShefferty, David
Akeroyd, Michael A.
author_facet Whitmer, William M.
McShefferty, David
Akeroyd, Michael A.
author_sort Whitmer, William M.
building Nottingham Research Data Repository
collection Online Access
description The most important parameter that affects the ability to hear and understand speech in the presence of background noise is the signal-to-noise ratio (SNR). Despite decades of research in speech intelligibility, it is not currently known how much improvement in SNR is needed to provide a meaningful benefit to someone. We propose that the underlying psychophysical basis to a meaningful benefit should be the just noticeable difference (JND) for SNR. The SNR JND was measured in a series of experiments using both adaptive and fixed-level procedures across participants of varying hearing ability. The results showed an average SNR JND of approximately 3 dB for sentences in same-spectrum noise. The role of the stimulus and link to intelligibility was examined by measuring speech-intelligibility psychometric functions and comparing the intelligibility JND estimated from those functions with measured SNR JNDs. Several experiments were then conducted to establish a just meaningful difference (JMD) for SNR. SNR changes that could induce intervention-seeking behaviour for an individual were measured with subjective scaling and report, using the same stimuli as the SNR JND experiment as pre- and post-benefit examples. The results across different rating and willingness-to-change tasks showed that the mean ratings increased near linearly with a change in SNR, but a change of at least 6 dB was necessary to reliably motivate participants to seek intervention. The magnitude of the JNDs and JMDs for speech-intelligibility benefits measured here suggest a gap between what is achievable and what is meaningful.
first_indexed 2025-11-14T19:35:57Z
format Article
id nottingham-38746
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T19:35:57Z
publishDate 2016
publisher Springer
recordtype eprints
repository_type Digital Repository
spelling nottingham-387462020-05-04T17:46:36Z https://eprints.nottingham.ac.uk/38746/ On detectable and meaningful speech-intelligibility benefits Whitmer, William M. McShefferty, David Akeroyd, Michael A. The most important parameter that affects the ability to hear and understand speech in the presence of background noise is the signal-to-noise ratio (SNR). Despite decades of research in speech intelligibility, it is not currently known how much improvement in SNR is needed to provide a meaningful benefit to someone. We propose that the underlying psychophysical basis to a meaningful benefit should be the just noticeable difference (JND) for SNR. The SNR JND was measured in a series of experiments using both adaptive and fixed-level procedures across participants of varying hearing ability. The results showed an average SNR JND of approximately 3 dB for sentences in same-spectrum noise. The role of the stimulus and link to intelligibility was examined by measuring speech-intelligibility psychometric functions and comparing the intelligibility JND estimated from those functions with measured SNR JNDs. Several experiments were then conducted to establish a just meaningful difference (JMD) for SNR. SNR changes that could induce intervention-seeking behaviour for an individual were measured with subjective scaling and report, using the same stimuli as the SNR JND experiment as pre- and post-benefit examples. The results across different rating and willingness-to-change tasks showed that the mean ratings increased near linearly with a change in SNR, but a change of at least 6 dB was necessary to reliably motivate participants to seek intervention. The magnitude of the JNDs and JMDs for speech-intelligibility benefits measured here suggest a gap between what is achievable and what is meaningful. Springer 2016-04-15 Article PeerReviewed Whitmer, William M., McShefferty, David and Akeroyd, Michael A. (2016) On detectable and meaningful speech-intelligibility benefits. Advances in Experimental Medicine and Biology, 894 . pp. 447-455. ISSN 0065-2598 Signal-to-noise ratio Just-noticeable difference Speech intelligibility Hearing impairment http://link.springer.com/chapter/10.1007%2F978-3-319-25474-6_47 doi:10.1007/978-3-319-25474-6_47 doi:10.1007/978-3-319-25474-6_47
spellingShingle Signal-to-noise ratio
Just-noticeable difference
Speech intelligibility
Hearing impairment
Whitmer, William M.
McShefferty, David
Akeroyd, Michael A.
On detectable and meaningful speech-intelligibility benefits
title On detectable and meaningful speech-intelligibility benefits
title_full On detectable and meaningful speech-intelligibility benefits
title_fullStr On detectable and meaningful speech-intelligibility benefits
title_full_unstemmed On detectable and meaningful speech-intelligibility benefits
title_short On detectable and meaningful speech-intelligibility benefits
title_sort on detectable and meaningful speech-intelligibility benefits
topic Signal-to-noise ratio
Just-noticeable difference
Speech intelligibility
Hearing impairment
url https://eprints.nottingham.ac.uk/38746/
https://eprints.nottingham.ac.uk/38746/
https://eprints.nottingham.ac.uk/38746/