Content uploaded by Christian Giguère
Author content
All content in this area was uploaded by Christian Giguère on Nov 07, 2015
Content may be subject to copyright.
Hearing Research Laboratory
Christian Giguère
Audiology/SLP Program
University of Ottawa, Canada
Elliott H. Berger
Division Scientist
3M, Indianapolis, U.S.A.
Exploration of Flat Hearing Protector
Attenuation and Sound Detection in Noise
168th Meeting ASA, Indianapolis, 27-31 October 2014
Context
•Flat-response HPDs have nearly-uniform attenuation
across frequency.
•They better preserve the spectral balance of sounds and
are often recommended when good speech
communication or sound perception is essential,
especially for users with high-frequency hearing loss.
•Few studies are available and relatively little is known
regarding when such devices can be beneficial (Casali et
al. 2010 for a review).
•Flat/uniform-attenuation HPDs are described in some
standards (e.g. EN 458; CSA Z94.2), but the definition of
“flatness” is not specified.
Context
•One exception is the German criterion for railway and
road traffic workers for good “SIGNAL AUDIBILITY (GENERAL),
SPEECH INTELLIGIBILITY AND PERCEPTION OF INFORMATIVE
OPERATING SOUND”(Liedkte, 2002, 2009)
•Based on Zwicker’s model, the critical slope is 3.6 dB/oct.
Fulfill(NO)
(slope≥3.6dB/oct)
Fulfill(YES)
(slope<3.6dB/oct)
Methods
Purpose:
•Study interaction between the noise spectrum, shape of
the HPD function and hearing loss on detection thresholds.
Design:
•Two noise spectra from NIOSH 100 database
•Two hearing configurations (normal and hearing loss)
•Four HPD attenuation sets
Effect of slope at a fixed protected level
Effect of protected level at a fixed slope
Etymotic®Musicians Earplugs™ and 3M™ HiFi™ earplugs
German guidelines (Fulfill NO and YES)
Outcome:
•Predicted detection thresholds using a psychoacoustic
model of warning sound perception in noise (Zheng et al.,
2007) based on Glasberg-Moore’s model.
Noises
NIOSH 100 Database:
•NIOSH(9,8): avg. of all spectra with large LC-LA= 8-9 dB
•NIOSH(3,2): avg. of all spectra with typical LC-LA= 2-3 dB
Hearing Configurations
Two hypothetical users:
•Normal Hearing (NH)
•Hearing loss (HLoss): Male, 55 yrs,
35 yrs @ 90 dBA, 10th percentile
Audiogram Auditoryfilters
1600Hz 3200Hz
NH
HLoss
ISO1999
HPD1
HPD2
Data Analysis
UNP
HPD1
HPD2
Threshold elevation IN NOISE (dB SPL)
HPD1
HPD2
Detection Thresholds (dB SPL) HEARINGLOSS
HPD2
HPD1
UNP
inNoise
NORMALHEARING
inQuiet
Results
Dataset 1: Selected slopes at a fixed protected level (75 dBA)
NIOSH(9,8) NIOSH(3,2)
8 dB/oct6 dB/oct0 dB/oct 4dB/oct2 dB/oct
Results
Dataset 1: Effect of slope at a fixed protected level (75 dBA)
NIOSH(9,8) NIOSH(3,2)
8 dB/oct6 dB/oct0 dB/oct 4dB/oct2 dB/oct
HEARINGLOSS
Upward spread of maskingAbsolute hearing threshold
Results
Dataset 2: Selected protected levels at a fixed slope (4 dB/oct)
NIOSH(9,8) NIOSH(3,2)
85 dBA
80 dBA75 dBA70 dBA60 dBA 65 dBA
Results
Dataset 2: Effect of protected level at fixed slope (4 dB/oct)
NIOSH(9,8) NIOSH(3,2)
HEARINGLOSS
Upward spread of maskingAbsolute hearing threshold
85 dBA
80 dBA75 dBA70 dBA60 dBA 65 dBA
Results
Dataset 3: Etymotic® Musicians Earplugs™
and 3M™ HiFi™ (ER20) earplugs
ER9 0.1
Mean
Slope
ER15 ‒ 0.3
ER20 1.4
ER25 1.3
NIOSH
(9,8)
NIOSH
(3,2)
dB/oct
84
81
77
71
dBA
85
82
77
72
dBA
Results
NIOSH(9,8) NIOSH(3,2)
HEARINGLOSS
Upward spread of maskingAbsolute hearing threshold
ER25
ER20ER15ER9
Dataset 3: Etymotic® Musicians Earplugs™
and 3M™ HiFi™ (ER20) earplugs
Results
Mean
Slope
NO 4.7
YES 2.3
NIOSH
(9,8)
NIOSH
(3,2)
dB/oct
73
67
dBA
70
68
dBA
Dataset 4: German guidelines (Liedkte 2002, 2009)
Fulfill
Criterion
Results
NIOSH(9,8) NIOSH(3,2)
HEARINGLOSS
Upward spread of maskingAbsolute hearing threshold
Fulfill criterion YES NO
Dataset 4: German guidelines (Liedkte 2002, 2009)
Conclusions
•Complex interaction between noise spectrum, hearing
status, HPD function and protected level on sound detection.
•For normal hearing, maximum effect is ≤ 2 dB even under
“extreme” conditions.
•For hearing loss, sound detection above 2500 Hz becomes
progressively more sensitive to the slope of the HPD function
and the protected level achieved.
•HPD slope not exceeding 2-3 dB/oct at a protected level not
lower than 75-80 dBA needed to control effects to less than
5 dB up to 5000 Hz for moderately-severe hearing loss.
•Further work to focus on quantifying benefits of flat HPDs in
more complex tasks involving speech communication, music
perception/playing, and recognition and interpretation of
important sounds to attend to in one’s environment.