International Journal of Audiology

ISSN: 1499-2027 (Print) 1708-8186 (Online) Journal homepage: https://www.tandfonline.com/loi/iija20

Spectral analysis of hearing protector impulsive

insertion loss

Cameron J. Fackler, Elliott H. Berger, William J. Murphy & Michael E. Stergar

To cite this article: Cameron J. Fackler, Elliott H. Berger, William J. Murphy & Michael E. Stergar
(2017) Spectral analysis of hearing protector impulsive insertion loss, International Journal of
Audiology, 56:sup1, 13-21, DOI: 10.1080/14992027.2016.1257869

To link to this article: https://doi.org/10.1080/14992027.2016.1257869

Published online: 25 Nov 2016.

Submit your article to this journal

Article views: 263

View related articles

View Crossmark data

Citing articles: 4 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=iija20
 International Journal of Audiology 2017; 56: S13–S21

Original Article

Spectral analysis of hearing protector impulsive insertion loss

Cameron J. Fackler1, Elliott H. Berger1, William J. Murphy2 & Michael E. Stergar1

1
3M Personal Safety Division, Indianapolis Acoustical Laboratories, Indianapolis, IN, USA and 2Hearing Loss Prevention Team, National
Institute for Occupational Safety and Health, Cincinnati, OH, USA

Abstract
Objective: To characterise the performance of hearing protection devices (HPDs) in impulsive-noise conditions and to compare various
protection metrics between impulsive and steady-state noise sources with different characteristics. Design: HPDs were measured per the
impulsive test methods of ANSI/ASA S12.42-2010. Protectors were measured with impulses generated by both an acoustic shock tube and
an AR-15 rifle. The measured data were analysed for impulse peak insertion loss (IPIL) and impulsive spectral insertion loss (ISIL). These
impulsive measurements were compared to insertion loss measured with steady-state noise and with real-ear attenuation at threshold
(REAT). Study sample: Tested HPDs included a foam earplug, a level-dependent earplug and an electronic sound-restoration earmuff.
Results: IPIL for a given protector varied between measurements with the two impulse noise sources, but ISIL agreed between the two
sources. The level-dependent earplug demonstrated level-dependent effects both in IPIL and ISIL. Steady-state insertion loss and REAT
measurements tended to provide a conservative estimate of the impulsively-measured attenuation. Conclusions: Measurements of IPIL
depend strongly on the source used to measure them, especially for HPDs with less attenuation at low frequencies. ISIL provides an
alternative measurement of impulse protection and appears to be a more complete description of an HPD’s performance.

Key Words: Hearing protection; impulse noise; insertion loss; spectral analysis

Introduction These include dimensional requirements, a requirement for heated

Impulsive noises are short-duration, high-level sounds. They are flesh simulations in the ear canals and circumaural areas, and a
created by many sound sources, including firearms, impact ham- minimum requirement for self-insertion loss, which is the maximum
mers, some pneumatic tools and automotive airbags. An important attenuation that could be measured with the ATF if its artificial ears
question and area of research is how effective hearing protectors are were completely sealed. At least two commercially-available ATFs
at protecting users from impulsive noises. have been designed to meet these requirements; measurements
The current American National Standard for objectively reported in this paper were obtained with an ATF manufactured by
measuring the insertion loss of hearing protection devices (HPDs) the French-German Research Institute of Saint-Louis (ISL).
is ANSI/ASA S12.42-2010. The impulsive-noise portion of that Historically, several studies have measured the impulse-noise
document standardises a measurement of impulse peak insertion performance of various HPDs. Impulsive noises are commonly
loss (IPIL). This is the amount in decibels (dB) that an HPD reduces characterised by their peak sound pressure level and A duration,
the peak sound pressure level of an impulse at the HPD user’s ear. defined as the time required for the initial pressure wave to rise to
This measurement requires the use of an acoustic test fixture (ATF), its peak pressure and return momentarily to ambient (Coles et al,
since a key component is the measurement of an unoccluded ear’s 1968). Ylikoski et al (1987, 1995) measured the impulse peak
response to impulses with peak levels as high as 168 dB. reduction of various hearing protectors for impulses generated by a
An ATF is essentially an artificial head, containing ear variety of weapons, using microphone-in-real-ear (MIRE) tech-
simulators mimicking human ears and microphones to record the niques in actual shooting conditions. They found that attenuation
sound pressure that would be present at the location of the eardrum. was greatest for small-caliber weapons such as pistols and least for
To ensure that the attenuation measured with an ATF is reasonably large-caliber weapons such as howitzers and cannons. Zera and
representative of actual use data, ANSI/ASA S12.42-2010 defines Mlynski (2007) also studied the peak reduction of impulses by
requirements for ATFs used for impulsive noise measurements. earmuffs, using a laboratory shock tube impulse source and an ATF

Correspondence: Cameron J. Fackler, 3M Personal Safety Division, Indianapolis Acoustical Laboratories, 7911 Zionsville Road, Indianapolis, IN 46268, USA. E-mail:
[email protected]

(Received 27 October 2016; accepted 2 November 2016)

ISSN 1499-2027 print/ISSN 1708-8186 online ß 2016 British Society of Audiology, International Society of Audiology, and Nordic Audiological Society
DOI: 10.1080/14992027.2016.1257869
S14 C. J. Fackler et al

Abbreviations longer bypassing the inherent passive attenuation of the earmuff.

ANSI American National Standards Institute; Lastly, the 3MÔ Combat ArmsÔ Single-Ended Earplug contains a
ATF acoustic test fixture; passive acoustic filter that responds nonlinearly to increasing sound
B&K Brüel & Kjær; pressure levels, such that higher-level impulses are attenuated more
HPD hearing protection device; than lower-level incident noises. This device, subsequently referred
IPIL impulse peak insertion loss; to as the level-dependent (LD) earplug, also contains a switch that
ISIL impulsive spectral insertion loss; allows the filter to be opened or closed. The device was tested in the
ISL French-German Research Institute of Saint-Louis; open mode to quantify its level-dependent attenuation.
LD level-dependent; To complete an impulse-noise characterisation following the
NI National Instruments; methods of ANSI/ASA S12.42-2010, an impulse-noise source was
REAT real-ear attenuation at threshold; used to generate test signals that were measured by an ATF and
SD standard deviation free-field microphone. The outputs of the ATF ear-simulator
microphones, with and without an HPD fitted to the ATF, were
recorded with the impulse-noise source generating impulses over a
for measurements. Their measurements were made at peak levels range of test levels. At each test level, 6 open-ear calibration
waveforms and 10 occluded-ear test waveforms were measured.
from 150 to 170 dB, with impulses at the higher levels having a
Testing at a given level began and ended with the measurement of
shorter A duration. They found that as the impulse A duration
three calibration waveforms. Before each test waveform measure-
decreases, the reduction in peak level by the earmuff increases.
ment, a sample of the HPD under test was fitted to the ATF. ANSI/
Giguère and Kunov (1989) constructed an ATF with enhanced
ASA S12.42-2010 stipulates that at least five samples of the hearing
self-insertion loss. Among other measurements, they used it to
protector under test are to be measured; the 10 test waveforms at
assess the complex (amplitude and phase), frequency-dependent
each level correspond to two repeat fittings of each HPD sample.
insertion loss of HPDs exposed to gunshot impulses. They
Five samples each of the LD earplug and electronic earmuff were
recommended that future work supplement the frequency-based
measured. A fresh sample was used for each of the 10 measure-
measurements by including time-domain impulse measurements,
ments at each test level for the foam earplug.
such as the reduction of peak sound pressure level.
To study the influence of the impulse-noise source on the
Recently, a number of studies have been conducted with the goal
resulting HPD measurements, impulses were generated by two
of measuring IPIL, as standardised in ANSI/ASA S12.42-2010, for
sources with different characteristics. The acoustic shock tube and
a variety of hearing protectors, using a variety of ATFs and impulse
horn [shown in the left of Figure 2; constructed following the design
sources. Murphy et al (2012) studied four hearing protectors,
outlined in Khan et al (2012) with a catenoidal horn] in the Impulse
measuring IPIL with a single-eared ATF, using an AR-15 as the
Laboratory of 3M’s Indianapolis Acoustical Laboratories generated
impulse source. Murphy et al (2015) used a shock tube impulse
impulses with A durations of approximately 1.0–1.5 ms. A Rock
source and measured IPIL for four hearing protector conditions,
River LAR-15LH LEF-T Tactical Operator-L rifle firing a Federal
including double protection.
American Eagle .223 Remington cartridge with 55 grains of powder
This work combines aspects of several of the aforementioned
and a full metal jacket bullet (referred to as the AR-15 rifle for the
studies. Using both an acoustic shock tube and an AR-15 as impulse
remainder of this paper), shown in the right of Figure 2, generated
sources, we measured both IPIL and the frequency-dependent
impulses with A durations of approximately 0.3–0.5 ms; measure-
(spectral) insertion loss of several HPDs. Measurements were
ments were taken outdoors in Lawrence County, IN, USA.
conducted with the ISL ATF mentioned above, including measure-
Although representative of impulses that may be encountered by
ments of insertion loss using a steady-state pink noise signal.
users of HPDs while shooting small firearms, these AR-15 impulses
Additionally, the steady-state and impulsive insertion loss meas-
did not meet the requirements of ANSI/ASA S12.42-2010, which
urements were compared to real-ear attenuation at threshold
calls for test impulses with A durations between 0.5 and 2.0 ms.
(REAT) measurements, which use human listeners to characterise
Examples of typical impulse waveforms and their corresponding
the attenuation of the same HPDs.
one-third-octave-band spectra generated by the two impulse sources
are shown in Figure 3. The spectra were calculated by filtering 10
representative free-field waveforms from each impulse source at
Methods
each test level into one-third-octave bands, computing the equiva-
Hearing protector impulse noise measurements lent sound level in each band and averaging these values across the
Three HPDs (shown in Figure 1) were characterised in impulse 10 impulses. Each spectrum has been normalised to its average
noise conditions using the methods of ANSI/ASA S12.42-2010. The value across frequency from 1000 to 8000 Hz, to facilitate
hearing protectors studied include the 3MÔ E-A-RÔ ClassicÔ comparison between the various impulses, especially in the low
Earplug, a roll-down foam earplug that is a passive device whose frequencies. As would be expected by comparing the A durations, it
attenuation is not expected to vary with the ambient noise level. The is clear from these data that the impulses generated by the shock
3MÔ PELTORÔ TacticalProÔ Electronic Headset (referred to as tube contain more low-frequency energy than the impulses
the electronic earmuff in the remainder of this paper) contains generated by the AR-15. Additionally, it is evident that for a
environmental microphones and level-dependent electronics that given source, lower-level impulses contain more low-frequency
provide amplification of (allowing a wearer to hear) low-level energy.
sounds. This device was tested with the electronics powered on and To generate an impulse with the shock tube, a polyester film
the volume control adjusted to its middle setting, providing membrane is clamped between flanges at one end of a cylindrical
approximately unity gain. At higher sound pressure levels, the pressure chamber. After the chamber is pressurised, the membrane
electronics cease to amplify the environmental sounds, thus no is ruptured with a pointed lance and the compressed air is rapidly
 Impulsive spectral insertion loss S15

Figure 1. Measured HPDs. 3MÔ E-A-RÔ ClassicÔ Earplug (foam earplug, left), 3MÔ Combat ArmsÔ Single-Ended Earplug (level-
dependent earplug, centre) and 3MÔ PELTORÔ TacticalProÔ Electronic Headset (electronic earmuff, right).

Figure 2. Experimental setups. Left: Acoustic shock tube and horn with ATF and blast probe microphone positioned in front of the horn in
a hemi-anechoic chamber. Right: AR-15 with ATF and free-field microphone positioned for measurements at the 150 dB peak level.

released, creating a shock wave and impulse after propagation down The ATF and blast probe microphones were located approximately
the shock tube and attached acoustic horn. The peak level of the 1 m away from the mouth of the acoustic horn, with the blast probe
resulting impulse is controlled by varying the membrane thickness located on the central axis of the horn and the right ear of the ATF
and chamber air pressure. located approximately 0.32 m to the left of the blast probe. Both the
For the shock-tube measurements reported in this paper, an ISL blast probe and ATF were aligned to point into the throat of the
ATF was used to measure the open- and protected-ear waveforms acoustic horn to capture the impulses in grazing incidence, with
and a GRAS Type 67SB blast probe measured the free-field their microphone elements located approximately 1.1 m above the
waveforms. The ISL ATF contained GRAS Model RA0045-S5 ear floor of the test room. Fibreglass panels were placed on the floor
simulators with GRAS Type 40BP 1/4-inch pressure microphones, between the mouth of the horn and the blast probe and ATF
GRAS Type 26AC microphone preamplifiers and ear canals microphones to minimise acoustic reflections. To generate impulses
approximately 18 mm long (to the microphone protection grid). with a nominal peak level of 168 dB, membranes with a thickness of
The blast probe contained a GRAS Type 40DP 1/8-inch pressure approximately 51 mm (2 mil) were ruptured at a gauge pressure of
microphone and GRAS Type 26AC microphone preamplifier. approximately 310 kPa (45 psi). For impulses with nominal peak
S16 C. J. Fackler et al

Figure 3. Typical free-field impulse waveforms and nominal peak levels (left) and corresponding spectra (right) for measurements with the
shock tube and AR-15 impulse sources. Impulse waveforms are shown as a 10 ms detailed view in the region of the impulse peak. Shock
tube impulse waveforms are presented in the left column, with AR-15 impulses in the middle column. The corresponding spectra computed
from the complete 105 ms waveforms are presented on the right. Each spectrum has been normalised to its average value from 1000 to
8000 Hz.

levels of 150 dB, membranes approximately 25 mm (1 mil) thick The noticeable dip in the 150 dB AR-15 spectrum at approximately
were ruptured at a gauge pressure of approximately 68 kPa (10 psi). 200 Hz may be attributed to the presence of this ground reflection.
Due to limitations in the generation of low-level impulses by the At the 132 dB test level, the ground reflection is merged with the
shock tube setup, the third set of impulses was measured with initial impulse, causing a slight elongation of the impulse’s
nominal peak levels of 140 dB, not the 132 dB called for in ANSI/ A duration.
ASA S12.42-2010. To generate impulses of this level, membranes For both the shock tube and AR-15, measurements were
with a thickness of approximately 13 mm (0.5 mil) were ruptured at recorded with a National Instruments (NI) PXI-1033 data acquisi-
a gauge pressure of approximately 35 kPa (5 psi). tion chassis with an NI PXI-4462 data acquisition module,
For impulses generated with the AR-15 rifle, the impulse peak controlled by Trident software (Nelson Acoustics, Elgin, TX,
level was controlled by varying the distance from the rifle’s muzzle USA). The microphones were polarised to 200 V with GRAS Type
to the location of the ATF and free-field microphone. The ISL ATF 12AA power modules. The power module was set to a unity gain of
described above measured the in-ear waveforms, while a Brüel & +0 dB for the free-field microphone channel at all test levels. The
Kjær (B&K) Type 4136 1/4-inch pressure microphone and B&K gain on the ATF channels was adjusted to 20 dB at the highest test
Type 2669 microphone preamplifier located beside the ATF level (168 dB peak), unity gain of +0 dB for testing at 150 dB and a
measured the free-field waveforms. To capture impulses with a gain of +20 dB for testing at the lowest levels (140 dB for shock
nominal peak level of 168 dB, the ATF and free-field microphone tube tests, 132 dB with the AR-15). Microphone sensitivities were
were positioned approximately 1.8 m to the side of and slightly verified with a GRAS Type 42AP pistonphone at the start of each
behind the AR-15’s muzzle. For impulses with a nominal peak level day of testing and anytime gain settings were adjusted or
of 150 dB, the ATF and microphone were located approximately instrumentation was moved. The microphone signals were sampled
4.3 m behind the rifle’s muzzle and they were positioned approxi- at a rate of 102.4 kHz and a resolution of 24 bits; the resulting three-
mately 25.7 m behind the muzzle to capture impulses with a channel waveforms (one free-field and two ATF ear channels) were
nominal peak level of 132 dB. The extended distance required for saved digitally as calibrated WAV files for later processing.
the 132 dB impulses required the ATF and free-field microphone to
be positioned within approximately 10 m of trees at the edge of the
range. In all cases, the rifle’s muzzle and the ATF and free-field Data analysis
microphone diaphragms were located approximately 1.5 m above After the waveforms described in the previous section had been
the ground surface, with the free-field microphone oriented for measured, the recorded waveforms were processed to determine the
grazing incidence with respect to the impulse source. This produced impulsive performance of the HPDs under test. In the following
ground reflections which are clearly seen in the AR-15 impulses analysis procedures, the waveforms consisted of the 24-bit digital
shown in Figure 3, especially at the 150 and 168 dB test levels. recordings described above. Prior to analysis, the peak sound
 Impulsive spectral insertion loss S17

pressure of the free-field channel of each recording was used to frequency-dependence, we also analysed the impulsive waveforms
align and truncate the waveforms; a total of 105 ms of recorded data for spectral insertion loss using the following procedure.
is used for the analysis of each trial, beginning 5 ms before the free- Transfer functions to estimate the open-ear ATF response
field peak. from the free-field microphone response were calculated as in
Equation (1) above. For each occluded-ear test measurement
pclosed, level, ear, k, m ðtÞ, the appropriate transfer function (Hlevel, ear ðf Þ)
IMPULSE PEAK INSERTION LOSS was used to estimate the corresponding open-ear waveform
For reference, a summary of the IPIL data analysis methods from popen, est, level, ear, k, m ðtÞ, as in Equation (2) above. A one-third-
ANSI/ASA S12.42-2010 is provided in the following. octave-band spectrum from 125 to 8000 Hz was then computed
The six open-ear calibration measurements at each test level for each estimated open-ear waveform and measured occluded-ear
were used to compute a frequency-domain transfer function (Hðf Þ) waveform. The spectrum value at each band was calculated as the
between the free-field microphone response (represented by the equivalent level (Leq) after filtering the waveform into the
recorded waveforms pFF, cal ðtÞ) and the open-ear fixture response corresponding one-third-octave band.
(from the waveforms popen ðtÞ). The calculation was performed as For each occluded measurement, the measured occluded-ear
  one-third-octave-band spectrum was arithmetically subtracted from
1X 6
F popen, level, ear, n ðtÞ the corresponding estimated open-ear spectrum. This resulted in the
Hlevel, ear ð f Þ ¼ , ð1Þ
6 n¼1 F ðpFF, cal, level, n ðtÞÞ one-third-octave-band insertion loss spectrum, from 125 to 8000 Hz.
ISIL was computed by averaging the insertion loss spectra across
where n is the calibration trial number, F denotes the discrete-time the 10 measurement trials at each test level, ultimately yielding
Fourier transform and the subscripts level and ear indicate that three ISIL curves (one at each test level) for each HPD under test.
unique transfer functions are computed at each test level for both
ATF ears.
These transfer functions were used to estimate the open-ear ATF STEADY-STATE INSERTION LOSS AND REAL-EAR ATTENUATION
response (popen, est ) that would have been measured during each of The insertion loss of each HPD under test was also measured in
the 10 occluded-ear measurements at each test level. Given the free- steady-state (continuous) noise conditions. These measurements were
field microphone response (waveform pFF, test ðtÞ) from each taken with the same ISL ATF described above for the impulse test
occluded-ear test, the ATF ear responses were estimated as methods. In this case, the ATF was positioned in a quasi-diffuse pink-
noise sound field with a free-field sound pressure level of 120 dB at
popen, est, level, ear, k, m ðtÞ
ð2Þ the head-centre position, measured without the ATF present.
¼ F 1 ðHlevel, ear ð f Þ  F ðpFF, test, level, k, m ðtÞÞÞ, With no HPD fitted to the ATF, three open-ear measurements of
pink noise were recorded and stored as one-third-octave-band sound
where k is the sample number (from 1 to 5) of the HPD under test, m
pressure levels. Each HPD was fitted to the ATF three times and
is the repeat fitting number (1 or 2) of the HPD and F 1 denotes the
occluded-ear measurements of the one-third-octave-band sound
discrete-time inverse Fourier transform. As before, the subscripts
pressure levels under the HPD were recorded for each fit. The
level and ear indicate that the unique transfer function Hlevel, ear ðf Þ is
insertion loss was obtained by averaging the occluded-ear meas-
used for each test level and fixture ear to predict the open-ear
urements (across three trials and both ATF ears) and subtracting
response.
these values from similarly-averaged open-ear levels. Like the ISIL
For each test measurement, the IPIL of the HPD was computed
described above, these measurements were obtained at all one-third-
from the estimated open-ear and measured occluded-ear
octave bands with centre frequencies from 125 to 8000 Hz.
waveforms as
REAT measurements of the HPDs under test were also obtained.
   ! These data were measured according to the experimenter-fit
maxt abs popen, est, level, ear, k, m ðtÞ
IPILlevel, ear, k, m ¼ 20 lg    procedures of ANSI S3.19-1974. In contrast to the steady-state
maxt abs pclosed, level, ear, k, m ðtÞ measurements conducted at 120 dB and the impulsive measure-
ð3Þ ments obtained at higher peak sound pressure levels, REAT
quantifies the insertion loss of an HPD at sound pressure levels
where maxt denotes the maximum value across all time samples (of corresponding to the threshold of hearing of human participants.
the absolute value of the respective waveform) and lg denotes the Even in the protected condition, the sound pressure levels of the
common logarithm. Performing the calculation for each ATF ear, diffuse sound field did not exceed approximately 70 dB.
HPD sample k and repeat fitting m results in 20 IPIL values at each Each trial during a REAT measurement consisted of measuring
test level. Before reporting the data, the average and standard the test participant’s auditory threshold with and without an HPD
deviation (SD) were computed across the IPIL values at each test fitted. Each threshold was determined as the participant listened to
level for every HPD tested. one-third-octave-band noise stimuli and a complete trial included
measurements with a range of one-third-octave bands at centre
frequencies from 125 to 8000 Hz. The insertion loss was calculated
IMPULSIVE SPECTRAL INSERTION LOSS by subtracting the open-ear threshold level from the occluded-ear
The IPIL analysis described in the previous section quantifies the threshold level. Three repeat fittings of each HPD under test were
impulsive protection of an HPD with a single number based on measured on each of 10 participants. The REAT insertion loss
the instantaneous reduction in maximum peak level between the average and standard deviation were computed across these 30 trials
unoccluded and occluded time waveforms measured. This single for each HPD.
number does not explicitly quantify any frequency-dependence During REAT measurements, the active electronics of the
of the HPD’s attenuation properties. To characterise the earmuff were turned off. These electronics amplify low-level
S18 C. J. Fackler et al

sounds, which would obscure the occluded thresholds measured on the REAT measurements on human participants were also in
the test participants. Steady-state insertion loss and impulsive tests generally good agreement with the other insertion loss values,
were conducted with the electronics turned on, as the product would although they tended to be slightly offset towards less attenuation
be used in these conditions. The level-dependent electronics limit than the other measurements.
the output of the device in high-level sounds; at the steady-state and
impulsive levels reported here, any electronic reproduction of
ambient sounds was effectively disabled. Electronic earmuff
Results for the electronic earmuff are shown in Figure 5. Although
this earmuff contains active circuitry to amplify and reproduce low-
Results
level environmental sounds, at the impulsive and steady-state test
For simplicity, the three impulse peak test levels are labelled as levels, these electronics should have been effectively turned off.
nominally 132, 150 and 168 dB in the following figures and table. Thus, the earmuff should have responded like a passive, non-
Shock-tube testing at the lowest level was conducted at a nominal level-dependent HPD. However, the measured IPIL increased as a
peak level of 140 dB in order to produce more ideal impulse function of impulse peak test level. Similar results have been
waveforms, as described above. observed in previous impulse measurements of passive earmuff
As required by ANSI/ASA S12.42-2010, the ISL ATF has a self- peak attenuation (e.g., Ylikoski et al, 1987, 1995; Zera & Mlynski,
insertion loss of greater than or equal to 60 dB at all test frequencies, 2007; and unpublished results from our own laboratory). As in these
exceeding the bone-conduction limits of human participants. earlier studies, the increase in IPIL with increasing test level may
Consequently, it is possible for insertion loss measured with the have been attributed to a corresponding decrease in both the
ATF to exceed values that are physically achievable by humans. impulse’s A duration and its low-frequency energy content.
Since one of the research goals of this study was to compare the The difference in IPIL between the shock-tube and AR-15
impulsively-measured results to measurements obtained on human measurements may be explained in a similar manner. The
participants, the ATF spectral insertion-loss measurements were attenuation of the electronic earmuff decreases at lower frequencies,
adjusted to account for bone conduction and physiological noise particularly below approximately 500 Hz. Since the shock-tube
masking following the methods of Schroeter and Poesselt (1986). impulses have more acoustic energy at these frequencies, the IPIL
Unfortunately, such a correction does not exist for impulsive peak measured with a shock tube may have been expected to be lower
measurements. than that measured with the AR-15 impulses.
IPIL results for all three HPDs studied are listed in Table 1. The Unlike the IPIL measurements, the ISIL measured for the
results are presented for both impulse sources and all three test electronic earmuff increased only slightly, if at all, with increasing
levels. They are presented as the mean and standard deviation of test level. This was in agreement with previous findings that
IPIL across the 10 trials for each device at each test level. These earmuffs have a linear response to shock waves below approxi-
results are also provided in graphical form for each protector in the mately 150–160 dB and do not appreciably demonstrate nonlinear
following sections. effects until over 170 dB (Zera & Mlynski, 2007). Further evidence
for the linear, level-independent earmuff response was provided by
the steady-state insertion loss and REAT measurements, which
Foam earplug agreed nicely with the ISIL data.
Results measured for the foam earplug are pictured in Figure 4. One ISIL curve that appeared out of place was that measured
Considering first the IPIL, it is seen that the performance is with the AR-15 at 132 dB. For the outdoor measurements taken with
relatively constant across impulse peak levels, with an IPIL of the AR-15, some trees and other potential reflecting surfaces were
approximately 50 dB at all three test levels. Since this earplug is a located in the vicinity of the ATF and free-field microphone at the
passive device with no intentional level-dependent properties, this 132 dB level, as previously noted. Although the active electronics
agreement was expected. There was also good agreement between would not reproduce the peak of the impulse, they may have
the IPIL measurements taken with the shock tube and AR-15 amplified and reproduced any low-level reflected or reverberant
impulse sources. energy following the impulse peak. This would tend to decrease the
Considering the spectral insertion loss measurements, similar measured ISIL, without significantly affecting the peak insertion
agreement is noted between both impulse sources and also between loss, since ISIL was computed using energy integrated over the
impulsive and steady-state results. The insertion loss measured with entire waveform recording. Such an effect would not have been
steady-state pink noise was in very good agreement with that noticed with the other HPDs studied, since they are both passive
measured with both impulse sources, at all test levels. Additionally, devices.

Table 1. Measured IPIL mean values and standard deviations (mean ±1 SD) in dB.
Impulse peak level
Device Impulse source 132 dB 150 dB 168 dB
Foam earplug Shock tube 43.3 ± 2.3 50.3 ± 4.0 55.4 ± 3.9
AR-15 49.7 ± 1.8 49.6 ± 4.2 51.2 ± 1.8
Electronic earmuff Shock tube 21.4 ± 2.2 28.3 ± 3.5 36.7 ± 2.9
AR-15 31.4 ± 1.3 41.2 ± 2.1 44.2 ± 1.3
Level-dependent earplug Shock tube 12.5 ± 1.5 20.6 ± 1.0 35.7 ± 1.5
AR-15 24.3 ± 0.7 29.8 ± 1.7 36.8 ± 2.0
 Impulsive spectral insertion loss S19

Level-dependent earplug from 24 dB (at the 132 dB test level), to 30 dB (at 150 dB), to 37 dB
Results for the LD earplug are shown in Figure 6. Unlike the foam (at 168 dB).
earplug and electronic earmuff, the LD earplug contains a passive Considering the ISIL, a similar increase in attenuation with
acoustic filter intended to create a level-dependent attenuation by increasing test level was evident, indicative of the inherently LD
responding nonlinearly to increasing sound pressure level (Hamery attenuation of this HPD. Unlike the IPIL data, which did not agree in
& Dancer, 1998). Therefore, the IPIL was expected to increase with magnitude between shock tube and AR-15 measurements, the ISIL
increasing impulse level above a transition level of approximately data appeared to agree well between the two impulse sources, varying
120 dB (Allen & Berger, 1990), as was seen in the measured results. only as a function of test level. Since the spectral insertion loss may
As measured with the shock tube, IPIL increased from approxi- be considered a more broadly-applicable attenuation property of an
mately 13 dB (at an impulse peak level of 140 dB), to 21 dB (at HPD than its response to an arbitrary time waveform, this insertion-
150 dB), to 36 dB (at 168 dB). The AR-15 IPIL values increased loss agreement between impulse sources was expected.

Figure 4. Measured IPIL (left) and spectral insertion loss (right) for the foam earplug. Impulsive values measured with the shock tube
source are presented as solid bars and lines, while impulsive values measured with the AR-15 source are shown with hashed bars and dashed
lines. IPIL and REAT values are shown with ±1 SD error bars. As noted in the text, the lowest-level shock tube impulses had a nominal peak
level of 140 dB.

Figure 5. Measured IPIL (left) and spectral insertion loss (right) for the electronic earmuff. Impulsive values measured with the shock tube
source are presented as solid bars and lines, while impulsive values measured with the AR-15 source are shown with hashed bars and dashed
lines. IPIL and REAT values are shown with ±1 SD error bars. As noted in the text, the lowest-level shock tube impulses had a nominal peak
level of 140 dB.
S20 C. J. Fackler et al

Figure 6. Measured IPIL (left) and spectral insertion loss (right) for the level-dependent earplug. Impulsive values measured with the
shock tube source are presented as solid bars and lines, while impulsive values measured with the AR-15 source are shown with hashed bars
and dashed lines. IPIL and REAT values are shown with ±1 SD error bars. As noted in the text, the lowest-level shock tube impulses had a
nominal peak level of 140 dB.

Examining the frequency-dependence of the insertion loss, it is Although the data collection and IPIL analysis procedure
seen that for the LD earplug, the insertion loss was greatest around employed in this work follow the methods of ANSI/ASA S12.42-
3000–4000 Hz and decreases at lower frequencies. As with the 2010, the impulses generated by the AR-15 rifle did not meet the
electronic earmuff, the lower IPIL measured with the shock tube A-duration requirements of that standard. These impulses had an
may be explained by the fact that the shock-tube impulses have A duration of approximately 0.3–0.5 ms; S12.42-2010 requires
greater low-frequency energy than the AR-15 impulses. In addition A durations of at least 0.5 ms and not more than 2 ms. As we have
to the inherently LD response of this earplug, the decrease in low- shown, IPIL measured with longer-A-duration impulses typically
frequency energy of the impulses at higher test levels also likely underestimates that which would be measured with a shorter
contributed to the measured increase in IPIL. However, as can be A duration. Future revisions to the S12.42 standard should consider
seen with the ISIL curves, the attenuation becomes flatter across allowing the use of shorter A durations for impulsive measurements.
frequencies at higher levels (the low-frequency attenuation However, IPIL measured with impulses meeting the current
increases more than the high-frequency attenuation). This may A-duration requirements would provide a conservative estimate of
help to explain why the shock tube and AR-15 IPIL measurements the peak reduction of an HPD worn for protection from impulses
almost agree at the 168 dB test level. generated by weapons such as rifles and pistols, including the
As with the previous HPDs, the steady-state insertion loss AR-15 used in this study.
measured with 120 dB pink noise and the REAT insertion loss As noted above, there is currently no defined method to adjust
measured on human participants demonstrated good agreement with measured IPIL values to account for the bone-conduction limits to
the lowest-level ISIL. The REAT values tended to provide a attenuation encountered with human participants. A potential for
conservative estimate of the lowest level of impulsive protection. future research would be to develop such a method, so that IPIL
Since the acoustic filter element in the LD earplug was not expected measured on an ATF might be more predictive of the performance
to demonstrate appreciable nonlinear effects until levels over obtainable by a human wearer of the HPD tested.
approximately 120 dB, the agreement between steady-state and
lowest-level-impulsive measurements was expected and indicated
linear performance of the device at low levels. Conclusions
The attenuation properties of three hearing protectors, including a
passive foam earplug, an electronic earmuff and a passive level-
Discussion
dependent earplug, were measured in impulsive and steady-state
When considering the impulsive measurements of an HPD, the ISIL conditions. Measurements include IPIL, ISIL, steady-state spectral
appears to be a more complete description of the device’s insertion loss and REAT. For the impulsive measurements, both a
performance than IPIL. In all cases where the HPD under test shock tube and an AR-15 were used as impulse noise sources,
was responding linearly to the applied test signals, spectral insertion generating impulses with varying amounts of low-frequency energy.
loss measured with either steady-state or impulsive noise was in IPIL was found to depend on the spectrum of the test impulse as
agreement. Additionally, in these cases, the REAT measurements well as the attenuation characteristics of the HPD under test. For
provided a conservative estimate of the steady-state and even the HPDs with less attenuation at low frequencies, IPIL was lower for
lowest-level-ISIL measurements. the shock tube impulse source, because it produced impulses
 Impulsive spectral insertion loss S21

containing more low-frequency energy. IPIL for these devices also trademarks of 3M Company, used under licence in Canada.
increased with increasing test level; in this study, lower-level PELTOR is a trademark of 3M Svenska AB, used under licence
impulses also contained more low-frequency energy than higher- in Canada.
level impulses. For the LD earplug, increases in IPIL with
increasing test level were also observed, as expected from the
presence of the nonlinear acoustic filter element in the earplug. References
Spectral insertion loss for an HPD under test, whether measured
Allen, C.H. & Berger, E.H. 1990. Development of a unique passive hearing
with the shock tube or AR-15 impulsive sources or steady-state
protector with level-dependent and flat attenuation characteristics. Noise
noise, agreed between test methods at a given test level. For Cont Eng J, 34, 97–105.
protectors that may be considered level-independent (including the ANSI S3.19-1974. American National Standard Methods for the
passive foam earplug, the electronic earmuff above approximately Measurement of Real-Ear Protection of Hearing Protectors and
120 dB and the LD earplug below approximately 120 dB), the Physical Attenuation of Earmuffs. New York: American National
spectral insertion loss results also agreed well between test levels. Standards Institute.
For the LD earplug, insertion loss increased as a function of test ANSI/ASA S12.42-2010. American National Standard Methods for the
level above approximately 120 dB, as expected. Furthermore, Measurement of Insertion Loss of Hearing Protection Devices in
shock-tube and AR-15 ISIL measurements were in agreement at Continuous or Impulsive Noise Using Microphone-in-Real-Ear or
a given impulse test level, demonstrating that the frequency- Acoustic Test Fixture Procedures. New York: American National
Standards Institute.
dependent insertion loss is a more robust description of the HPD
Coles, R.R.A., Garinther, G.R., Hodge, D.C. & Rice, C.G. 1968. Hazardous
than its time-domain response to arbitrary time waveforms, which is exposure to impulse noise. J Acoust Soc Am, 43, 336–343.
used for the computation of IPIL. For all HPDs studied over the Giguère, C. & Kunov, H. 1989. An acoustic head simulator for hearing
range of test levels studied in this paper (up to approximately protector evaluation. II: Measurements in steady-state and impulse noise
170 dB SPL), REAT provided a conservative estimate of the environments. J Acoust Soc Am, 85, 1197–1205.
spectral insertion loss, measured with either steady-state or Hamery, P.J.F. & Dancer, A.L. 1998. New nonlinear earplugs for protection
impulsive noise. against impulse noise. In: Carter, N. & Soames, R.F. (eds.) Noise Effects
’98 – Proceedings of the 7th International Congress on Noise as a
Public Health Problem; Sydney, Australia: Noise Effects ’98 PTY LTD,
Acknowledgements pp. 95–97.
Khan, A., Murphy, W.J. & Zechmann, E.L. 2012. Design and Construction
The authors thank Amir Khan and Taichi Murata from the National
of an Acoustic Shock Tube for Generating High-Level Impulses to Test
Institute for Occupational Safety and Health for their assistance Hearing Protection Devices. EPHB Report No. 250-12a. Cincinnati:
with data collection for this study. The authors also thank Jeff National Institute for Occupational Safety and Health.
Hamer for the use of his shooting range and his assistance with data Murphy, W.J., Fackler, C.J., Berger, E.H., Shaw, P.B. & Stergar, M. 2015.
collection. Portions of this research were presented at the 2016 Measurement of impulse peak insertion loss from two acoustic test
Annual Conference of the National Hearing Conservation fixtures and four hearing protector conditions with an acoustic shock
Association in San Diego, California. tube. Noise Health, 17, 364–373.
Murphy, W.J., Flamme, G.A., Meinke, D.K., Sondergaard, J., Finan, D.S.,
et al. 2012. Measurement of impulse peak insertion loss for four hearing
Disclaimer protection devices in field conditions. Int J Audiol, 51, S31–S42.
Schroeter, J. & Poesselt, C. 1986. The use of acoustical test fixture for the
The findings and conclusions in this report are those of the authors measurement of hearing protector attenuation. Part II: Modeling the
and do not represent any official policy of the Centers for Disease external ear, simulating bone conduction, and comparing test fixture and
Control and Prevention or the National Institute for Occupational real-ear data. J Acoust Soc Am, 80, 505–527.
Safety and Health. Mention of company names and products does Ylikoski, J., Pekkarinen, J. & Starck, J. 1987. The efficiency
not constitute endorsement by the CDC or NIOSH. of earmuffs against impulse noise from firearms. Scand Audiol, 16,
85–88.
Ylikoski, M.E., Pekkarinen, J.O., Starck, J.P., Pääkkönen, R.J. & Ylikoski,
Declaration of interest: Cameron Fackler, Elliott Berger and J.S. 1995. Physical characteristics of gunfire impulse noise and its
Michael Stergar are employed by 3M, the manufacturer of the attenuation by hearing protectors. Scand Audiol, 24, 3–11.
hearing protectors studied in this work. 3M, E-A-R, Classic, Zera, J. & Mlynski, R. 2007. Attenuation of high-level impulses by earmuffs.
Combat Arms, TacticalPro and the colour yellow for earplugs are J Acoust Soc Am, 122, 2082–2096.