Modeling the interaction between the hearing protector attenuation function and the hearing loss profile on sound detection in noise

Abstract

Few studies document the exact conditions when flat/uniform hearing protectors can be beneficial in the noisy workplace. This modeling study reports on the interaction between the user’s hearing loss profile and the shape and amount of the attenuation function on sound detection thresholds in noise. For normal-hearing users, detection thresholds are found to be hardly affected by use of hearing protectors, even in extreme conditions of low-frequency noise and steeply sloping attenuation functions. With aging and noise-induced hearing loss, sound detection above about 2000 Hz becomes progressively more sensitive to the slope of the attenuation function as well as to the overall protected level achieved. Shallower slopes may be warranted for users with hearing loss to limit the upward spread of masking in low-frequency noise, while controlling the total amount of attenuation at high frequencies prevents excessive elevation of absolute thresholds. Decisions regarding hearing protector selection also entail consideration of the principal auditory tasks that are anticipated and the important sounds to which a worker may need to attend.

Full text

Modeling the interaction between the hearing
protector attenuation function and the hearing
loss profile on sound detection in noise
Christian Giguère
Audiology/SLP Program, University of Ottawa, 451 Smyth Rd., Ottawa, Canada K1H 8M5.
Elliott H. Berger
Personal Safety Division, 3M, 7911 Zionsville Rd., Indianapolis, IN 46268, U.S.A.
Summary
Few studies document the exact conditions when flat/uniform hearing protectors can be beneficial
LQWKHQRLV\ZRUNSODFH7KLVPRGHOLQJ VWXG\UHSRUWVRQWKHLQWHUDFWLRQEHWZHHQ WKHXVHU¶VKHDULQJ
loss profile and the shape and amount of the attenuation function on sound detection thresholds in
noise. For normal-hearing users, detection thresholds are found to be hardly affected by use of
hearing protectors, even in extreme conditions of low-frequency noise and steeply sloping
attenuation functions. With aging and noise-induced hearing loss, sound detection above about
2000 Hz becomes progressively more sensitive to the slope of the attenuation function as well as
to the overall protected level achieved. Shallower slopes may be warranted for users with hearing
loss to limit the upward spread of masking in low-frequency noise, while controlling the total
amount of attenuation at high frequencies prevents excessive elevation of absolute thresholds.
Decisions regarding hearing protector selection also entail consideration of the principal auditory
tasks that are anticipated and the important sounds to which a worker may need to attend.
PACS no. 43.50.Hg, 43.66.Vt
1. Introduction
1
Hearing protector devices (HPDs) with flat or
nearly uniform attenuation across frequencies
preserve the spectral balance of workplace sounds
and are often recommended when, in addition to
protection, good signal audibility, speech
communication and auditory awareness are
essential [1]. Such HPDs may be especially
indicated for users with high-frequency hearing
loss to maintain audibility at all frequencies. Still,
few studies are available on the exact conditions
when the intended benefits of flat/uniform devices
would arise [2]. Furthermore, while flat/uniform
HPDs are described in several acoustical standards
and/or national documents [3-5], the definition of
³IODWQHVV´ LV often unspecified or given broad
tolerances.
In [5], a general criterion based on the
slope of the linear regression of the mean
attenuation values between 125 and 4000 Hz is
specified. Based on prior research with road traffic
and railway workers, HPDs with an attenuation
slope less than 3.6 dB/octave are deemed to fulfill
UHTXLUHPHQWV IRU ³VLJQDO DXGLELOLW\ JHQHUDO
speech intelligibility, and the perception of
operative sounds´ In a proposed revision to
standard EN458:2004 [4], flat devices are defined
more simply as HPDs with H minus L attenuation
values less than or equal to 9 dB. In addition to
flatness characteristics, the assumed protection
value of a HPD is also an important parameter to
considerer in some standards [3,4] in order to
reduce the risks of overprotection. HPDs selected
to achieve a protected level between 5 and 10 dBA
below the national regulation level are deemed to
achieve a good protection outcome.
It is unclear if such simple guidelines for
flatness and protection level achieved can find
general use given, among other factors, the wide
range of spectral characteristics of workplace
noises and the large spread of possible hearing
profiles in the workforce. The goal of this
modeling study is to gain more insight into the
complex interaction between the hearing protector
attenuation function, the noise spectrum, and the
hearing status of the worker.
Copyright© (2015) by EAA-NAG-ABA V , ISSN 2226-5147
All rights reserved
1967

EuroNoise 2015
31 May - 3 June, Maastricht
C. Giguère et al.: Modeling the...
1968

2.4. Simulation and data analysis
Pure-tone masked detection thresholds from 125 to
8000 Hz were computed in the two noises and for
the three hypothetical workers using the
algorithms specified in the psychoacoustic model
in [6]. This was carried out both for unprotected
listening and when protected according to the
attenuation functions from the two HPD datasets.
Unprotected detection thresholds were then
subtracted from protected detection thresholds to
highlight cases of threshold elevation from the use
of HPDs. Thus, the unprotected detection
thresholds served as a baseline to quantify the
effects of hearing protection for each individual
worker and noise. Such a threshold elevation can
occur due to an overly large loss of hearing
sensitivity from the combined effects of the
absolute hearing thresholds and the amount of
HPD attenuation at one or more frequencies
(referred to here as Case 1 elevation) and/or due to
an increase in the upward spread of masking from
the interaction between the noise spectrum, the
broadening of the auditory filters and the shape of
the HPD function (Case 2 elevation).
3. Results
Figure 4 shows the detection threshold elevation
(dB) arising from the attenuation functions of the
HPD Dataset 1, separately for the three workers.
The results are for NIOSH (9,8), the noise with a
large LC-LA. For Worker 1, the maximum
threshold elevation was less than 2 dB across the
entire range of frequencies and for all HPD
attenuation slopes simulated. The maximum effect,
occurring at 400 Hz, is related to a peculiarity in
the NIOSH (9,8) spectrum which shows a sudden
drop of noise energy near that frequency. This, in
turn, led to some upward spread of masking from
the lower frequency energy that passed more easily
through the HPDs with the largest slopes of
increasing attenuation with frequency (4-8
dB/octave).
For Worker 2, there were also minimal
effects over the entire range of frequencies, but
only up to a slope of about 4 dB/octave (Figure 4).
At steeper slopes of 6 and 8 dB/octave, the
detection thresholds progressively increased over
unprotected thresholds for frequencies above
3000-4000 Hz. This was almost equally the result
of Case 1 and Case 2 threshold elevation effects.
For Worker 3, only the flat attenuation function of
0 dB/octave produced minimal effects over the
entire range of frequencies. HPD Functions with
slopes of 2 dB/octave or greater progressively
incurred higher threshold elevations above 2000
Hz. Case 2 threshold elevation was the controlling
factor for the two largest slopes.
Results for the NIOSH (3,2) noise with the
HPD Dataset 1 (not shown) were similar to
NIOSH (9,8), except that the lower LC-LA value
and the less rich low frequency spectrum (Figure
1) produced somewhat less upward spread of
masking. There were no threshold elevation effects
at 400 Hz for any of the three workers with this
noise.
Figure 5 shows the detection threshold
elevation (dB) arising from the attenuation
functions of the HPD Dataset 2 applied to noise
Figure 2. HPD Dataset 1: Attenuation functions for
various slopes (dB/octave) at a fixed protected level of
75 dBA in NIOSH (9,8).
Figure 3. HPD Dataset 2: Attenuation functions for
various protected levels (dBA) at a fixed slope of 4
dB/octave in NIOSH (9,8).
EuroNoise 2015
31 May - 3 June, Maastricht
C. Giguère et al.: Modeling the...
1969

EuroNoise 2015
31 May - 3 June, Maastricht
C. Giguère et al.: Modeling the...
1970

4. Discussion and conclusions
This computational study highlights the complex
nature of the interaction between the hearing loss
profile of the user, the HPD attenuation function
on sound detection in noise, and to some extent the
noise spectrum.
For normal-hearing users, the maximum
effect was less than 2 dB even under the most
extreme condition investigated of a low-frequency
noise (large LC-LA value between 8-9 dB) and use
of a HPD with a very steep attenuation curve (8
dB/octave). Furthermore, sound detection in noise
appears quite insensitive to the overall protected
level achieved for normal-hearing users, down to
at least 60 dBA. These results indicate that, for
sound detection in noise, there may be little value
in placing restrictions on attenuation slope and
protected level achieved for this class of user.
Such a finding corroborates earlier reports that
auditory perception in noise is hardly affected by
use of HPDs for normal-hearing users [1].
By contrast, for users with hearing loss,
detection thresholds in some noise situations may
be affected by use of hearing protectors with
steeply sloped attenuation functions and/or
excessive attenuation. The negative effect may be
due to elevated absolute hearing thresholds arising
from a too-large HPD attenuation in the high
frequencies in conjunction with the hearing loss
(Case 1 elevation) and/or from an upward spread
of masking due to steep HPD attenuation functions
in the presence of either broadened auditory filters
or low-frequency noise (Case 2 elevation).
From Figures 4 and 5, the effects of HPDs
on sound detection in noise are found to be quite
GHSHQGHQW RQ WKH XVHU¶V KHDULQJ SURILOH DQG WKH
signal frequency. This finding presents a special
challenge when setting general guidelines for HPD
selection applicable to all users, as shown below.
Assuming the maximum detection threshold
elevation in noise to be not larger than 5 dB at
4000 Hz, to minimize adverse effects on the
speech frequency range for example, then a fairly
wide prescription consisting of an attenuation
slope of no more than 6-7 dB/octave and a
protected level not lower than 60-65 dBA may be
suitable for a user with mild hearing loss (W2).
Using the same criterion, a more stringent
prescription consisting of a slope not exceeding 3-
4 dB/octave and a protected level not less than 75
dBA may be required for a user with moderate-
severe high frequency hearing loss (W3). On the
other hand, if good sound detection is required at
frequencies greater than 4000 Hz, more stringent
requirements may be needed for HPD selection. It
is also important to note that workers W2 and W3
specified in this study are only two of a myriad of
possible case studies. In general, there are much
wider variations in hearing profiles across users
than variations in attenuation across available HPD
products, which greatly compounds to the problem
of setting general selection guidelines.
Simulations were also carried out with two
additional datasets of HPD functions, as reported
in [10]. To investigate the effects of purposely-flat
real products, the mean attenuation values of the
Etymotic® 0XVLFLDQV (DUSOXJVŒ DQG WKH 0Œ
+L)LŒ(DUSOXJZHUHFRQVLGHUHG7KHGHYLFHVKDYH
mean slopes varying from -0.3 to 1.4 dB/octave in
the range from 125 to 4000 Hz and, when applied
to NIOSH (9,8) and NIOSH (3,2), yield protected
values in the range 71-85 dBA. Not surprisingly,
these products led to reduced threshold-elevation
effects that were comparable to the shallowest
slopes investigated in Figure 4. Even for the
worker with the most hearing loss (W3), the
maximum predicted threshold elevation was 5 dB
or less up to 5000 Hz for all products in both
NIOSH (9,8) and NIOSH (3,2) noises. Simulations
were also carried out using the average attenuation
curves reported in [5] for groups of real hearing
protectors fulfilling and not fulfilling the German
criterion of a mean slope of less than 3.6
dB/octave. Note that these curves were found to
possess mean slopes of about 2.3 dB/octave
(fulfill) and 4.7 dB/octave (not fulfill). For the
normal-hearing worker (W1), the maximum
predicted threshold elevation was less than 2 dB
up to 8000 Hz for both groups of protectors in
either noises, similar to results for Datasets 1 and 2
(Figures 4-5), and the difference between the two
groups was at most 1 dB. For worker 3, the
threshold elevation due to the upward spread of
masking (Case 2) was consistent with that
presented in Figure 4 for each group of protectors
at comparable slopes, and thus favored the group
fulfilling the criterion. However, the controlling
factor for the threshold elevation was the loss of
hearing sensitivity (Case 1) and, owing to the
nearly equal amounts of attenuation they provide
at mid-to-high frequencies, both groups of
protectors led to the same outcome above 3000 Hz.
In practice, decisions regarding HPD
selection for hearing-impaired users will entail
some knowledge about the principal auditory tasks
EuroNoise 2015
31 May - 3 June, Maastricht
C. Giguère et al.: Modeling the...
1971

to carry out in the given workplace. Proper
consideration of the characteristics of acoustic
signals to attend to is also important. With regards
to the perception and design of audible danger
signals, for example, ISO 7731:2003 [11] specifies
that ³in the case of persons wearing hearing
protection or having a hearing loss, sufficient
signal energy should be present in the frequency
range below 1500 Hz.´ Such a recommendation is
clearly in line with the results of this
computational study. Below 2000 Hz, sound
detection in noise was found to be quite insensitive
to the HPD attenuation function over a wide range
of noise conditions and types of users from normal
hearing to moderate-severe hearing losses due to
aging and noise (Figures 4 and 5). Thus, warning
signal perception may be little affected by HPDs
for a wide range of users when proper
consideration is first given to the design of
warning sounds in the workplace. On the other
hand, there are many instances of incidental
sounds over which we have little or no control,
such a malfunctioning machine or parts falling
from a conveyor belt [1]. In these cases, hearing-
impaired users may benefit from shallower
attenuation slopes and reduced amounts of
attenuation, especially in cases where such sounds
are likely to generate mostly high-frequency
acoustic energy.
Finally, it is important to note that the
present study assumed continuous noises in the 95-
96 dBA range. In practice, workplace noises may
fluctuate in level over the course of the day. In
quieter periods, the effect of HPDs may be more
pronounced. Also, due to medical conditions and
other reasons, some workers may have hearing loss
configurations that depart from the typical profiles
due to aging and noise-induced hearing loss
investigated in the present study. Sound detection
in noise is also only one of many possible auditory
dimensions where flat/uniform hearing protectors
may have an impact. A recent computational study
also showed that speech perception may also be
little affected by the HPD attenuation function
over a wide range of situations, except for users
with a substantial amount of hearing loss [12].
Still, flat/uniform hearing protectors preserve the
spectral balance of sounds and they may provide
substantial benefits in terms of user acceptance
resulting from improved sound quality and
auditory situational awareness such as the
recognition, interpretation, or in the case of
entertainment-sounds such as music, the
enjoyment RI LPSRUWDQW VRXQGV LQ RQH¶V
environment. Further work is needed in this area.
References
[1] E.H. Berger: Hearing protection devices. Chapter 10
In: E.H. Berger, L.H. Royster, J.D. Royster, D.P.
Driscoll and M. Layne (eds.). The Noise Manual, Fifth
Edition, American Industrial Hygiene Association,
Fairfax VA, 2003, 379-453.
[2] J.G. Casali: Passive augmentations in hearing
protection technology circa 2010 including flat-
attenuatin, passive level-dependent, passive wave
resonance, passive adjustable attenuation, adjustable-
fit devices: Review of design, testing and research.
International Journal of Sound and Vibration 15
(2010) 187-195.
[3] CSA Z94.2-14: Hearing protection devices -
Performance, selection, care, and use. Canadian
Standards Association, 2014.
[4] EN 458: Hearing protectors - Recommendations for
selection, use, care and maintenance ± Guidance
document. European Standard, 2004.
[5] M. Liedkte: Specifying a general criterion for hearing
protectors with the aim of ensuring good acoustic
perception. International Journal of Occupational
Safety and Ergonomics 15 (2009) 163-174.
[6] Y. Zheng, C. Giguère, C. Laroche, C. Sabourin, A.
Gagné, M. Elyea: A psychoacoustic model for spec-
ifying the level and spectrum of acoustic warning
signals in the workplace. Journal of Occupational and
Environmental Hygiene 4 (2007) 87-98.
[7] ISO 1999: Acoustics ± Estimation of noise-induced
hearing loss. International Organization for
Standardization, 2013.
[8] C. Laroche, B. Josserand, H. Tran Quoc, R. Hétu, B.R.
Glasberg: Frequency selectivity in workers with noise-
induced hearing loss. Hearing Research 64 (1992) 61-
72.
[9] ANSI/ASA S12.68: Methods of Estimating Effective
A-Weighted Sound Pressure Levels When Hearing
Protectors are Worn. Acoustical Society of America,
2007 (R2012).
[10] C. Giguère, E.H. Berger: Exploration of flat hearing
protector attenuation and sound detection in noise.
168th Meeting of the Acoustical Society of America,
Indianapolis IN, 27-31 October 2014.
[11] ISO 7731: Ergonomics ² Danger signals for public
and work areas ² Auditory danger signals.
International Organization for Standardization, 2003.
[12] C. Giguère, E.H. Berger: Investigation of Hearing
Protector Attenuation Function on Sound Detection
and Speech Recognition in Noise. 40th Annual
Conference of the National Hearing Conservation
Association, New Orleans LA, 19-21 February 2015.
3M and HiFi are trademarks of 3M Company, used under
licence in Canada.
EuroNoise 2015
31 May - 3 June, Maastricht
C. Giguère et al.: Modeling the...
1972