| Article Access Statistics|
| Viewed||1617 |
| Printed||88 |
| Emailed||0 |
| PDF Downloaded||7 |
| Comments ||[Add] |
|Year : 2012
: 14 | Issue : 58 | Page
|Attenuation of peak sound pressure levels of shooting noise by hearing protective earmuffs
Paolo Lenzuni1, Tommaso Sangiorgi2, Luigi Cerini3
1 Department of Florence, Italian National Workers' Compensation Authority (INAIL), Florence, Italy
2 Department of Florence, Joint University of Florence-Careggi Hospital Work Unit, Florence, Italy
3 Department of Occupational Hygiene, Italian National Workers' Compensation Authority (INAIL), Rome, Italy
Click here for correspondence address
|Date of Web Publication||15-Jun-2012|
Transmission losses (TL) to highly impulsive signals generated by three firearms have been measured for two ear muffs, using both a head and torso simulator and a miniature microphone located at the ear canal entrance (MIRE technique). Peak SPL TL have been found to be well approximated by 40 ms short-L eq TL. This has allowed the use of transmissibilities and correction factors for bone conduction and physiological masking appropriate for continuous noise, for the calculation of REAT-type peak insertion losses (IL). Results indicate that peak IL can be well predicted by estimates based on one-third octave band 40 ms short L eq and manufacturer-declared (nominal) IL measured for continuous noise according to test standards. Such predictions tend to be more accurate at the high end of the range, while they are less reliable when the attenuation is lower. A user-friendly simplified prediction algorithm has also been developed, which only requires nominal IL and one-third octave sound exposure level spectra. Separate predictions are possible for IL in direct and diffuse sound fields, albeit with higher uncertainties, due to the smaller number of experimental data comprising the two separate datasets on which such predictions are based.
Keywords: Attenuation, hearing protection, impulse noise, peak SPL
|How to cite this article:|
Lenzuni P, Sangiorgi T, Cerini L. Attenuation of peak sound pressure levels of shooting noise by hearing protective earmuffs. Noise Health 2012;14:91-9
| Introduction|| |
The 2003/10/CE directive  stipulates that two quantities should be used for the assessment of occupational exposure to noise: the A-weighted daily noise exposure L EX and the C-weighted peak sound pressure level L p,C,peak . The same document specifies that compliance with exposure limit values (ELV) must be established for both descriptors taking "account of the attenuation provided by the individual hearing protectors worn by the worker0".
The attenuation of the peak SPL that hearing protector devices (HPDs) can provide has been the topic of numerous studies since Ward's (1968)  seminal study. Yet, it still lies outside the scope of HPD test standards, primarily ISO 4869-1:1990,  and no information on this topic in the form of manufacturer-declared values ISO 4869-2:1994  is available to the end user. Some guidance as to how to estimate the attenuation of the peak SPL is provided by EN 458:2004.  In the latter document, impulsive noises are classified as type 1, 2, or 3, depending on their spectral content being mostly concentrated at low (e.g., punch press), medium to high (e.g., nail gun), or high frequency (e.g., pistol). The attenuation of the peak is then estimated from the manufacturer-declared L, M, or H attenuation for continuous noise, respectively, which are calculated as outlined in ISO 4869-2:1994.  A further reduction of the attenuation by a fixed amount d m = 5 dB is required for type 1 and type 2 signals (no correction for type 3 signals).  This method leads to significant systematic underestimates (2 to 4 dB) with respect to measurements performed inside the ear canal, and its uncertainties are uncomfortably large, as much as 6 dB for type 1 and type 2 signals and about 4 dB for shooting noise. 
The performance of hearing protectors when exposed to impulsive noise has often been characterized in one-third octave bands. , However, the shockwave generated by highly impulsive fire arms signals that propagates through the HPD  implies that this information is of limited use in the estimate of the attenuation of the peak SPL.
A moderate correlation of the peak attenuation with the SNR (Single Number Rating, the simplest descriptor of the HPD attenuation) of the hearing protector  might alternatively be exploited to use SNR as a predictor of the peak IL. Unfortunately, given the large variability of data and the noise source peculiarities, where amplitude and A-duration (hence spectral shape) are strongly correlated, no general algorithm can be extracted to predict the peak attenuation to a specific impulsive signal by a specific HPD.
Among the sources of impulsive signals, firearms are outstanding in showing peak SPL so large (>155 dB) ,, to possibly imply noncompliance with the EC Directive peak ELV of 140 dB(C) even after the effect of the HPD is duly taken into account. Additionally, even if the ELV is not exceeded, peak levels in the 130 dB(C) range would limit to just a few the number of allowed daily events according to NIOSH recommendations.  Given the huge number of potentially exposed subjects, more than 20 millions in the United States alone,  an accurate estimate of the peak SPL of the protected ear is badly needed.
The objective of this article is to setup a simple prediction algorithm to allow a reliable estimate of the attenuation of the peak sound pressure level provided by an ear muff, in the specific case of shooting noise. In the ensuing sections of this article, first the experimental methods are described, then the data analysis is outlined, and finally the results are presented and discussed.
| Experimental Setup|| |
All tests were carried out in a six-lane firing range located in northern Italy. The shooting area outer walls are entirely covered with sound-absorbing materials. Barriers interposed between adjacent lanes consist of a ballistic steel central layer covered with chipboard, while the external layers are perforated thin steel panels filled with sound absorbing material.
Measurements were first performed on a Brüel & Kjær Head and Torso Simulator (HATS) type 4128-C,  positioned 1.5 m behind the shooter as shown in [Figure 1], and equipped with two Brüel & Kjær ear simulators, one at each ear Drum Reference Position (DRP), which include ½" microphones. The HATS line of sight was aligned with the source-receiver path. These measurements, whose lay-out is as shown in [Figure 2], will be referred to in this article as HATS tests. Measurements were then replicated using a MIRE technique, positioning one miniature Sennheiser MKE 2-5 Gold-C microphone near the entrance of each ear canal of the shooter [Figure 3]. These tests, whose lay-out is as shown in [Figure 4], will be referred to as MIRE tests.
The attenuation performance of hearing protectors is best quantified by measuring their insertion loss (IL). The IL is the arithmetic difference between the sound pressure levels measured at a fixed position in the ear with the HPD absent (unoccluded exposure) and with the HPD worn (occluded exposure), with other conditions identical. Unfortunately, such measurements turned out to be impossible in our case. For HATS tests, the DRP peak sound pressure level during unoccluded exposures was large enough to exceed the microphones' linearity range, leading to frequent overloads. For MIRE tests, unoccluded ear canal pressures would have been so large to pose a realistic threat to the subjects' hearing functionality, and such measurements had to be ruled out because of ethical considerations. Accordingly, for both HATS and MIRE tests the experimental quantity was the TL, the arithmetic difference between the sound pressure levels simultaneously measured at some distance from the ear and at a fixed position in the ear with the HPD worn. In our tests, external measurements were carried out using two Brüel & Kjær type 4135/A and 4135/B ¼" microphones (hereafter external microphones) positioned at a distance of 15 cm from the right and left ear, respectively. The low sensitivity of these microphones is best suited to withstand the very high sound pressures which occur in vicinity of a fire arm. IL were calculated using the procedure described below.
Ear muffs and descriptors of attenuation
In order to ensure a self-consistent framework for the application of the EC directive, the same descriptor of attenuation should be used for both L EX and L p,C,peak . The HPD attenuation of continuous noise is commonly estimated using manufacturer-declared octave-band REAT-type values obtained following the procedure outlined in ISO 4869-1.  Although based on hearing threshold measurements, which imply human subjective responses in audiometric tests, and not on sound pressure level measurements, these values are de facto used as if they were IL. We will hereafter refer to these values as "REAT-type Insertion Losses" or "Nominal Insertion Losses". Such values are shown in [Table 1] for the two ear muffs tested in this work. A REAT-type IL for the peak sound pressure level is therefore the quantity that we aim at calculating in this study.
Three weapons were used as sources of impulsive signals: a semi-automatic pistol, a submachine gun, and an assault rifle (hereafter pistol, gun, and rifle, respectively). One wavetrain (source: rifle, microphone: external) is shown in [Figure 5].
|Figure 5: Shooting event wavetrain, rifle-generated signal, and external microphone|
Click here to view
Four shots were fired for each weapon/muff combination. Each shooting event was acoustically characterized by the following two quantities:
Both quantities were measured simultaneously on the two internal and the two external microphones. Linear peak sound pressure levels were also measured.
- the C-weighted peak sound pressure level L p,C,peak ; C-weighted peaks rather linear peaks were targeted because L p,C,peak is the descriptor of exposure to highly impulsive noise selected by the 2003/10 EC directive;
- the 1/3 octave band spectrum of the short equivalent level integrated over a time range of N ms onward from the peak L eq,Nms .
| Data Analysis|| |
Experimental determination of the TL
Mean peak SPL measured by the external microphones is summarized in [Table 2]. Peak levels are all much larger in MIRE tests because of the proximity of external microphones to the signal source, 0.9 ± 0.2 m compared with 2.4 ± 0.2 m in HATS tests. Differences between left ear and right ear peak levels are always ≤1 dB, indicating excellent left-right symmetry. The TL of tested HPDs was measured between the external and the DRP microphones in HATS tests (TL HATS ) and between the external and the concha microphones in MIRE tests (TL MIRE ). Experimental determinations of the peak TL are summarized in [Table 3] for both HATS and MIRE tests.
|Table 2: Measured peak levels (dBC) in HATS and MIRE tests: external microphones|
Click here to view
MIRE tests produce relatively greater TL than HATS tests. This is in line with expectations, given the absence of the ear canal amplification associated to the position of the miniature microphone in the ear's concha [Figure 3]. The TL of rifle-generated signals is systematically lower than the other two. This is likely due to the larger fractional low-frequency (f < 1000 Hz) content of rifle-generated signals, coupled to the poorer performance of earmuffs in this range. HATS tests [left side of [Table 3] show some evidence of a systematic left-right difference, with a mean difference over the six pairs of 1.7 dB and an associated uncertainty of 0.9 dB. Applying a matched-pair t-test,  the resulting test statistics is t?=?1.78, while = 1.78, while t0.025,6 = 2.97. Accordingly, the left-right difference is not statistically significant at the P < 0.05 level. MIRE tests right side of [Table 3] exhibits an excellent left-right symmetry (mean difference 0.2 dB). Differences between C-weighted TL and linear TL were also measured and found to be consistently small (<1.8 dB in 9 of 12 cases). All conclusions reached in this work for C-weighted peak IL are therefore equally valid for linear weighting.
Short-L eq TL vs. peak TL
In order to translate measured TL into IL, several correction factors must be considered. All such factors, which will be discussed in the forthcoming sections, can be measured or extracted from literature studies for continuous sound and are usually estimated in one-third octave bands. Their application to peak SPL is not straightforward however, because of the unknown phase shifts that take place in the HATS ear canal, or in the human pinna, at different frequencies.
In order to test the possibility of applying the existing information for continuous sound to the peak SPL, the measured peak TL
has been compared with various short-L eq TL, integrated over N ms
[Table 4] shows the difference (TL peak - TL eq,Nms ) averaged over the two ears, for the six tested weapon/muff combinations and for N = 20, 40, 60, 100, and 240 (HATS tests).
|Table 4: TLpeak – TLeq,Nms (dB) difference for various weapon/ muff/integration time combinations (HATS tests)|
Click here to view
The rms difference presented in the last line is a good descriptor of overall agreement. The best match has been achieved for a short-L Aeq integration time of 40 ms. This can be understood considering that 40 ms is the shortest timescale that includes the majority of the event energy (20 ms is too short for this purpose). While one-third octave band filters are too slow to give an adequate response below 200 Hz with a 40-ms time base, this is largely irrelevant: in a very short burst, energy per Hz is roughly constant up to a cutoff frequency, so that in octave bands the majority of energy is concentrated well above 500 Hz.
The difference (TL peak - TL eq,40 ms ) is below 2 dB in four of six cases, and the root mean square is 2.4 dB. This is good evidence that differential phase shift is of minor significance, something not obvious given the sharply dropping phase response above 2000-3000 Hz.  [Table 5] shows the equivalent information for MIRE tests, leading to very similar conclusions.
The uncertainty on the difference TL peak - TL eq,Nms depends on the uncertainties on both TL peak and TL eq,Nms , the former providing the strongest contribution. We derive an estimate of u(TL peak ) from those cases for which we have at least four independent measurements of TL peak , obtaining a mean value u(TL peak ) = 0.6 dB. Because TL peak and TL eq,Nms are mutually correlated, the uncertainty on TL peak - TL eq,Nms is even smaller.  The conclusion that TL peak and TL eq,Nms show very good mutual agreement is therefore unaffected by considerations on associated uncertainties.
|Table 5: TLpeak – TLeq,Nms (dB) difference for various weapon/ muff/integration time combinations (MIRE tests)|
Click here to view
Conversion from measured TL to REAT-type IL
In order to convert the measured values of TL HATS and TL MIRE into the desired values of REAT-type IL, two separate correction factors are needed, which shall be discussed in the forthcoming subsections.
From TL to IL
The first correction factor translates TL into IL. For HATS tests, this has been expressed in the form of an "external-to-DRP" transmissibility TR EtD .
The function TR EtD (f) was calculated through custom-planned anechoic chamber experiments, where the setup has closely mimicked the shooting range lay out, that is, in the presence of the HATS/human head and using different microphones for DRP and external measurements (see the section on instrumentation). TR EtD (f) differs from the standard HATS transmissibility where the blank measurement is carried out at the HATS's head center position, in the absence of the HATS itself, and the same type of microphones are used for both external and internal measurements.
[Figure 6] shows TR EtD (f) compared with two HATS transmissibilities, one provided by Brüel & Kjær for the specific manikin used in this work and the other included in ISO 11904-2.  Although the three curves display the same qualitative profile, some quantitative differences show up: TR EtD displays a pronounced hump around 1500 Hz which is otherwise absent, and it also exhibits a much stronger high frequency spike between 10,000 and 20,000 Hz. Despite its size, the latter is largely irrelevant given the relatively low-energy content of our signals above 10,000 Hz.
|Figure 6: Transmissibilities for HATS tests: TREtD (solid), manufacturer TRHATS (dotted with filled symbols), ISO 11904-2 TRHATS (dashed with empty squares)|
Click here to view
A similar concept was applied to MIRE tests, resulting in an "external-to-concha" transmissibility TR EtC , as shown in [Figure 7]. Unlike TR EtD , no "benchmark" exists in this case to gauge the results of our measurements, due to the peculiar frequency response of the miniature microphone as well as the strong effects due to its position.
From HATS/MIRE to REAT
The second correction factor takes care of all the differences between HATS and MIRE tests on one side and REAT tests on the other. Such differences, which are extensively discussed in the literature, ,, can be traced to three separate mechanisms, bone conduction (BC), occlusion effects (OE), and physiological masking (PM). In this article, BC has been included making use of Berger's data , which also incorporate OE. PM has been taken instead from Schroeter and Poesselt. 
The cumulative effects of such mechanisms have been summarized by the quantity D BC+PM , i.e., the difference between sound pressure levels that take/do not take into account BC (including OE) and PM. Following the formulation suggested by Schroeter and Poesselt,  PM does not depend on the muff or the signal, while BC does. Accordingly, [Figure 8] shows the quantity D BC+PM for two muff/signal combinations. In the low frequency range, PM provides a significant contribution, whereas the impact of BC is negligible. Because the existence of PM implies higher REAT attenuations,  this leads to positive values of D BC+PM . At intermediate and high frequency, PM plays no role, while BC limits the maximum achievable attenuation in REAT experiments,  leading to large negative values of D BC+PM . It is also worth commenting on the systematically lower values found for the less performing H64FBV ear muff. This is likely due to its IL spectral shape, which allows a relatively larger transmission at frequencies below 1000 Hz compared with the other ear muff. Following the same line of reasoning used above, this results in lower values of D BC+PM .
|Figure 8: DBC+PM for two signal/muff combinations: pistol/H540A (black) and rifle/H64HBV (grey)|
Click here to view
Calculation of the peak IL
Thanks to the good match between TL eq,40 ms and TL peak that was previously established, the path leading from TL eq,40 ms to IL eq,40 ms , taking into account the appropriate transmissibility and the correction factors previously discussed, can be applied to the calculation of IL peak from TL peak .
First, one-third octave band 40 ms IL is calculated as
where TR(f) = TR EtD for HATS tests and TR(f) = TR EtC for MIRE tests.
The cumulative effect of transmissibility, BC and PM is then estimated as
where L' eq,40 ms (f) is 40 ms one-third octave band short-L eq measured under the HPD, and L ext,40 ms (f) is the same quantity measured by the relevant external microphone. In equation (4) the first term only includes the TL, while the second also includes the contributions of transmissibility, BC, and PM. Finally, this quantity is added to the experimental peak TL to give the peak IL
The symbol IL exp has been used to indicate that this quantity is entirely experimentally determined.
| Results and Applications|| |
Synthesis of results
[Table 6] shows the IL IL exp as calculated in equation (5), for all weapon/muff combinations. Such values represent our best estimate of the quantity to be used to take into account the muff's attenuation of the peak sound pressure level in assessing compliance with EC or NIOSH limits. The range of calculated IL (20-36 dB) is almost identical to the range found by Pääkkönen and Lehtomäki,  but this is largely coincidental, as different muffs can give widely differing values, and both samples are extremely small. [Table 6] also includes a second estimate of the IL
|Table 6: Experimental and nominal IL (dB) for various weapon/muff combinations|
Click here to view
calculated from 40 ms one-third octave band short L eq spectra measured by external microphones and nominal IL [Table 1]] interpolated in one-third octave bands.
[Table 6] shows that IL based on nominal attenuation values (equation 6) mostly overpredict those based on experimental measurements (equation 5). Distance between the two sets tends to vanish at the high end of the range. Correlation is good (R2 = 0.84) as also evident from [Figure 9].
|Figure 9: Correlation between ILexp and ILNOM. HATS data (squares) and MIRE data (circles)|
Click here to view
Differences [IL NOM - IL exp ] span the range [-2.3; +7.7] dB, mean = +3.1 dB, s = 2.6 dB. The positive sign of the mean is consistent with expectations, given the very artificial nature of the ISO 4869-1 testing procedure, which leads to well-known overestimates of the IL. Its small magnitude is, however, somewhat surprising, considering that the response of ear muffs is likely no longer linear above peak SPL ? 150 dB,  so that values measured at threshold should be looked upon with some scepticism. A possible explanation may have to do with the different tested sound fields direct in this work, largely diffuse in ISO 4869-1.  The attenuation of ear muffs in diffuse fields shows some tendency to be lower than in face-on direct sound fields,  which would partially offset the strong positive bias induced by the ISO 4869-1 testing procedure and shrink the observed differences. It should be stressed that because peak sound pressure levels are always associated to the first direct wave, the correct peak IL must be calculated in direct sound fields, so no correction factor accounting for the sound field has been introduced in equation (5).
A simplified predictive algorithm
The procedure presented in this article may be perceived as somewhat academic, mostly because it relies on experimental measurements of a quantity, the 40 ms short-L eq , which is largely outside the reach of occupational hygienists.
A simplified method, based on just two easily accessible quantities
has therefore been developed as a more practical tool.
- one-third octave band spectra of the external sound exposure level, SEL(f), replacing one-third octave band spectra of L eq,40 ms (f), to characterize the signal; the SEL is the sound pressure level that would result by compressing all the signal energy into a 1-second window;
- nominal IL IL NOM (f) [Table 1], to characterize the muff attenuation,
In [Figure 10], the simplified IL
|Figure 10: Correlation between ILexp and ILSEL. HATS data (squares) and MIRE data (circles)|
Click here to view
are compared with the more accurate estimates IL exp given by equation (5).
A simple linear regression gives
The two datasets present a fair correlation (R2 = 0.74). The simplified method provides a clear understimate of the IL when attenuation is low to moderate, while a good match exists for high attenuation values. Lower than expected values are the result of the higher low-frequency content of SEL, which integrates the full duration of the signal, compared with the L eq,40 ms . The larger energy fraction at lower frequencies translates into lower attenuations.
In order to assess the goodness of fit, estimates of uncertainties are required, which are not straightforward:
The resulting estimate is u(IL SEL ) ≈ 1.8 dB for HATS tests and 2.5 dB for MIRE tests. Because u(IL SEL ) is not negligible compared with u(IL peak ), the goodness of fit assessment requires that uncertainties on both axes are taken into account. The resulting value χn2 = 0.63 indicates that an acceptable description of experimental data has been achieved.
- IL SEL is basically a mean nominal IL weighted by the energy spectrum of the external SEL. The uncertainty is roughly given by the typical nominal standard deviation in relevant octave bands (where most of the impulse energy shows up) divided by the square root of the number of relevant octave bands. With typical standard deviations in the 2-3 dB range and four to six relevant bands, u(IL SEL ) ≈ 1 dB.
- The uncertainty on IL peak receives contributions from:
- The uncertainty on the experimental determination of the TL, which is mostly determined by the instrumental uncertainty for measurements of peak sound pressure levels. For HATS tests, this is presumably quite small, less than 1 dB. For MIRE tests, it is likely to be appreciably larger: a tentative value of 2 dB has been assumed.
- The uncertainty on transmissibility, which is probably significant given the peculiarities of our experimental set-up, and the fact that the main transmissibility peak occurs at frequencies where the signal energy is large. We have assumed 1.5 dB.
- The uncertainty on corrections for BC and PM is negligible in this context.
Possible sensitivity to the sound field
[Figure 10] shows that data for HATS and MIRE tests tend to occupy slightly different areas of the diagram, with IL NOM (MIRE) being systematically lower than IL NOM (HATS) for almost any weapon/muff combination.
Two independent linear fits to data for HATS and MIRE tests give
Correlation coefficients are R2 = 0.85 for equation (9a) and R2 = 0.91 for equation (9b). A possible explanation is that this difference reflects a difference in sound fields: in HATS tests, the manikin is located behind the barriers which separate adjacent shooting lanes; in MIRE tests on the opposite, the human head is positioned near the geometric centre of the area between two facing barriers. The sound field in the latter position is characterized by reflections and/or re-irradiation of sound from the barriers exposed to the direct wave, creating a more "reverberant" field compared with the roughly free field characterizing HATS tests. As previously discussed, the associated enhancement of the mid-low frequency component in MIRE tests compared with HATS tests implies a lower amplification in the human ear because the muff transmissibility is much lower below 1000 Hz than it is above [Figure 6]. Eventually this determines a lower IL for a given TL.
Following this line of reasoning, equations (9a) and (9b) can be tentatively associated to the peak IL in (roughly) direct and (roughly) diffuse sound fields, respectively. Equations (9a) and (9b) should be seen as somewhat speculative, given the very limited amount of available data on which they are based.
An alternative/additional possibility is that the microphone-to-source line of sight is partially occluded by the shooter's body for HATS tests, while it is interference-free for MIRE tests. The effective size of this "pseudo-barrier" is, however, very small, and it is unlikely to produce any measurable effect.
| Conclusions|| |
A study addressing nonshooting signals is currently in progress.
- The EC 2003/10 Directive requires that peak SPL values must not exceed the ELV of 140 dB(C) taking into account the hearing protector worn by the worker. Peak IL of hearing protectors are unfortunately not included in test standards and therefore not provided by manufacturers. This prevents any simple "on-paper" method to estimate attenuated peak levels, in analogy to methods used to infer attenuated continuous equivalent levels.
- Current methods for estimating the peak IL are too generic to give the accuracy needed for a reliable assessment of compliance/noncompliance with the 2003/10 Directive ELV.
- Impulsive signals generated by firearms represent the most obvious case in which a possible noncompliance should be assessed.
- Peak IL have been estimated in this work for earmuffs exposed to highly impulsive shooting-type signals (unprotected peak levels around 160 dB at the shooter's position). The estimate exploits the very good agreement found between peak and 40 ms rms TL in order to calculate REAT-type peak IL, which are experimentally inaccessible.
- A predictive algorithm has been developed which allows a reliable prediction of the peak IL using only manufacturer-declared attenuation values and measurements of SEL octave band spectra.
- Separate algorithms have been developed to predict IL in direct and diffuse fields. This should, however, be seen as somewhat speculative, given the limited amount of available data.
| References|| |
|1.||Directive 2003/10/EC of the European Parliament and of the Council, of 6 February 2003. On the minimum health and safety requirements regarding the exposure of workers to the risks arising from physical agents (noise) - (seventeenth individual directive within the meaning of article 16(1) of directive 89/391/eec), Official Journal of the European Union L42, p. 38-44, Brussels, Belgium, 2003. |
|2.||Ward WD. Proposed damage-risk criterion for impulse noise (gunfire). Report of Working Group 57. Washington DC, U.S.A: National Academy of Sciences - National Research Council Committee on Hearing, Bioacoustics and Biomechanics (CHABA); 1968. |
|3.||ISO. Acoustics - Hearing protectors - Part 1: Subjective method for the measurement of sound attenuation. Geneva, Switzerland: International Organization for Standardization; ISO 4869-1:1990. |
|4.||ISO. Acoustics - Hearing protectors - Part 2: Estimation of effective A-weighted sound pressure levels when hearing protectors are worn. Geneva, Switzerland: International Organization for Standardization; ISO 4869-2:1994. |
|5.||EN. Acoustics - Hearing Protectors - Recommendations for selection, use care and maintenance. European Committee for Standardization; EN 458. Brussels, Belgium; 2004. |
|6.||Smoorenburg GF. Assessment of hearing protector performance in impulsive noise. Final Report. TNO Report TM-96-C042. Soesterberg, Netherlands: TNO Human Factors Research Institute; 1996. |
|7.||Nakashima A, Buck K, Hamery P, De Mezzo S, Brom G. Impulse noise: measurement techniques and hearing protector performance. Defence R&D Canada, Technical Memorandum TM2006-231. Toronto, Canada; 2006. |
|8.||Murphy WJ, Tubbs RL. Assessment of noise exposure for indoor and outdoor firing ranges. J Occup Environ Hyg 2007;4:688-97. |
|9.||Zera J, Mlynski R. Attenuation of high-level impulses by earmuffs. J Acoust Soc Am 2007;122:2082-96. |
|10.||Pääkkönen R, Lehtomäki K, Myllyniemi J, Hämäläinen E, Savolainen S. Noise attenuation of hearing protectors in the human ear - a method description. Acust. and Acta Acust. 2000;86:477-80. |
|11.||NIOSH. Noise health hazard evaluation report: HETA-2002-0131-2898. National Institute for Occupational Safety and Health, Cincinnati, OH, U.S.A. |
|12.||Kardous CA, Murphy WJ. Noise control solutions for indoor firing ranges. Noise Control Eng J 2010;58:345-56. |
|13.||NIOSH. Comments of NIOSH on the OSHA ANPR Hearing conservation program for construction workers 29 CFR Part 1926, Docket H-011G. Cincinnati, OH, U.S.A: National Institute for Occupational Safety and Health;2002. |
|14.||Brüel & Kjær Head and Torso Simulator (HATS) - Type 4128C Product Sheet. Available from: http://www.bksv.com/products/telecomaudiosolutions/headtorso/headandtorsosimulatorhatstype4128c.aspx. [Last accessed on May 29, 2012]. |
|15.||Dowdy S, Wearden S. Statistics for Research. New York: Wiley; 1991. |
|16.||Mlynski R, Zera J. Measuring amplitude and phase response of earmuffs. Lille, France: In Proceedings of NoiseatWork conference; 2007. p. 1241-5. |
|17.||ISO/IEC Guide 98-3:2008 - Uncertainty of measurement-Part?3: Guide to the expression of uncertainty in measurement (GUM: 1995). Geneva, Switzerland. |
|18.||ISO. Acoustics - Determination of sound immission from sound sources placed close to the ear - Part 2: Technique using a manikin. Geneva, Switzerland: International Organization for Standardization. ISO 11904-2:2004. |
|19.||Casali JG, Mauney DW, Burks JA. Physical vs. Psychophysical measurement of hearing protector attenuation - a.k.a. MIRE vs. REAT. Sound Vib 1995;29:20-7. |
|20.||Berger EH. Preferred methods for measuring hearing protector attenuation. In Proceedings of Internoise 05, Rio de Janeiro, Brasil. Noise Control Foundation editor. Poughkeepsie, NY, U.S.A: 2005. p. 58-67 |
|21.||Reinfeldt S, Stenfelt S, Good T, Håkansson B. Examination of bone-conducted transmission from sound field excitation measured by thresholds, ear-canal sound pressure, and skull vibrations. J Acoust Soc Am 2007;121:1576-87. |
|22.||Berger EH. Laboratory attenuation of earmuffs and earplugs, both singly and in combination. Am Ind Hyg Assoc J 1983;44:321-9. |
|23.||Berger EH. Protection for infrasonic and ultrasonic noise exposure. EarLog14. Available from: http://www.e-a-r.com/pdf/hearingcons/earlog14.pdf. [Last accessed on 1984]. |
|24.||Schroeter J, Poesselt C. The use of acoustical test fixtures for the measurement of hearing protector attenuation. Part II: Modeling the external ear, simulating bone conduction, and comparing test fixture and real-ear data. J Acoust Soc Am 1986;80:505-27. |
|25.||Hagerman B, Olofsson A, Cheng J, Svensson E. Ear muff performance by threshold method and by probe microphone method. Acta Acust. 1995;3:569-74. |
|26.||Pääkkönen R, Lehtomäki K. Protection efficiency of hearing protectors against military noise from handheld weapons and vehicles. Noise Health 2005;7:11-20. |
|27.||Buck K. Performance of hearing protectors in impulse noise. RTO Lecture Series 219, Report RTO-ENP-11 AC/323(HFM)TP/31. In: Dancer AL, editor. Neuilly-sur-Seine, France: Res. And Tech. Org. (RTO) of NATO; 2000. p. 3-1 - 3-10. |
Italian National Workers' Compensation Authority (INAIL), Department of Florence, Via delle Porte Nuove 61, 50144 Florence
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9], [Figure 10]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]