Objective and subjective assessment of sound attenuation efficiency by individual hearing protectors with various acoustic filters – a preliminary study

Background : Hearing loss caused by excessive noise levels is one of the most common health risks for employees. One solution for noise reduction is the use of hearing protectors, which is one of the most effective methods for protecting hearing from noise at the workplace. In order to obtain different attenuation efficiency, individual hearing protectors can be equipped with a suitable acoustic filter. The effectiveness of the hearing protectors attenuation is based on real measurement of hearing thresholds for normal-hearing people with and without hearing protectors. However, this is a time-consuming process and the obtained values are characterized by quite large inter-individual variability. The optimal solution is to measure the attenuation characteristics based on the objective method (without the presence of the subject), the results of which will be in accordance with the results of subjective tests. Therefore, the main purpose of the research in this work was to measure the attenuation characteristics of individual hearing protectors with acoustic filters through the use of subjective and objective methods, and to compare the results in terms of the research methods. Methods : Measurements of the acoustic attenuation obtained by individual hearing protectors with designed F1, F2 and F3 acoustic filters, as well as full insert earplugs (without any acoustic filters) were carried out using two methods: objective and subjective. The objective measurements were carried out in an anechoic chamber. The artificial head (High-frequency Head and Torso Simulator Brül & Kjær Type 5128) was located at a distance of 3 m, directly opposite the loudspeaker. The test signal in the measurements was pink noise - in the frequency range up to 12.5 kHz and the level 85, 90 and 95 dB. The hearing protectors with and without acoustic filters were mounted in the Head and Torso Simulator which was connected with Pulse System Brül & Kjær. Five normal hearing subjects participated in the subjective measurements. A pink noise signal was used for one-third octave bands: 125, 250, 500, 1000, 2000, 4000 and 8000 Hz. The attenuation value was defined as the difference (in dB) between the hearing threshold of the test signal with a hearing protector and the hearing threshold determined without a hearing protector. Results : The results of the objective method proved that in both the full earplugs and the earplugs with the F1, F2 and F3 filters. In addition, the results of the objective method showed that in the whole frequency range the highest attenuation values are shown by the full earplugs, achieving slightly above 45 dB for frequency of 8 kHz. The attenuation values obtained from subjective measurements also confirmed that both the frequency and type of filter significantly affect the attenuation values of the tested hearing protectors. Unlike the results of the objective method, the subjective method did not indicate significant differences in attenuation when using F1 and F2 filters. Conclusions : The comparison of the average attenuation values obtained from the objective and subjective methods showed that in general the measurement method does not significantly affect the average attenuation values. In turn, the analysis of variance broken down into subgroups according to the types of filters used in the earplugs showed that the influence of the measurement method on the attenuation values is statistically significant when the F1 filter and full earplug are used. The results of this study partly confirmed the hypothesis that there is no significant impact of the measurement method on the attenuation characteristics of the earplugs with different types of acoustic filters. attenuation of passive and active hearing protectors based on an objective method, i.e. using the artificial head AFT 45CB and AFT (45CA). They examined three different types of passive hearing protectors (circum-aural, soft flange in ear and simple foam in-ear) and three types of active hearing protectors (supra-aural, circum-aural and one in-ear canal). They stated that the test results do not depend on the devices used (AFT 45 CB and AFT 45 CA).


Background
Noise is defined as any unwanted sound which can be annoying or unsafe to human health, and which may increase the risk of an accident at work. Noise can cause sleep disturbance, worsening of speech intelligibility, as well as temporary or permanent hearing loss. Hearing loss is a process that usually develops over years, because it is usually painless, gradual and practically imperceptible. The progress of hearing loss due to the chronic influence of noise depends on the level of noise exposure and its duration. Hearing loss caused by excessive noise levels is one of the most common health risks for employees. Over 20% of workers in Europe are at risk of permanent hearing loss. The safe and acceptable noise level, within an eighthour working day, is 85 dB. This noise level is often exceeded: a comparable or higher noise level (instantaneous values) occurs for example, along highways. Noise at the workplace can be limited by both organizational methods and technical means.
One solution for noise reduction is the use of hearing protectors, which is one of the most effective methods for protecting hearing from noise at the workplace. There are many types of hearing protectors: earmuffs, anti-noise earplugs and individual hearing protectors. Individual hearing protectors in the form of earplugs are made on the basis of the ear impression. Such earplugs, after insertion, tightly close the external auditory canal. In order to obtain different attenuation efficiency, individual hearing protectors can be equipped with a suitable acoustic filter. These are easily replaceable elements, which are selected depending on the acoustic conditions at the workplace. The selection of hearing protectors is carried out on the basis of measurements of sound pressure levels at the workplace. The choice of hearing protector means there is a need to determine the appropriate attenuation efficiencythis should not allow excessive acoustic protection. Such excessive protection could make it impossible to communicate with other people and hear warning sounds.
In the available literature, research topics concerning hearing protectors are mainly focused on the problem of how protectors influence the ability to attenuate sounds with different frequency characteristics and changes in time (broadband noise, bands of noise, sound pulses, gunfire) (Davis et al., 2011;Biabani et al., 2017;Fackler et al., 2017;Samelli et al., 2018).
In addition, the effect of hearing protectors on the ability to determine the location of a sound source (Zimpfer and Sarafian, 2014;Brown et al., 2015;Lee and Casali, 2017), changes in the level of speech intelligibility (Bockstael et al., 2011;Norin et al., 2011;Brown et al., 2015;Hiselius et al., 2015;Lee and Casali, 2017) as well as more complex aspects of sound sensations, e.g. changes in the perception of the sound of musical instruments by musicians using hearing protectors are also examined (Killion, 2012). It turns out that in the case of musicians, the sound levels can reach very high values, in the order of over 100 dB, which, with long-term exposure, may also result in hearing loss.
One of the problems resulting from the use of hearing protectors is the deterioration of speech intelligibility, especially when speech is presented against the background of noise, and when the level of speech sounds is low. Active hearing protectors that enable the selection of signal enhancement depending on the sound level reaching the user are a partial solution here (Bockstael et al., 2011;Brown et al., 2015;Lee and Casali, 2017). However, in one paper (Norin et al., 2011), the authors showed that using 3 different passive hearing protectors did not determine the impact of the type of the protector on speech intelligibility against noise, and showed that none of the hearing protectors significantly affected speech intelligibility. The one significant factor influencing speech intelligibility was the signal-to-noise ratio (SNR).
In the case of active hearing protectors (signal amplification function depending on the level of the input signal) a significant improvement in speech intelligibility was found, especially for lower SNR values (Hiselius et al., 2015).
In the studies determining the effectiveness of the attenuation provided by hearing protectors, the key issue is the research methodology according to ISO 4869-1, the main advantage of which is real measurement of hearing thresholds for normalhearing people with and without hearing protectors.
Comparison of hearing thresholds allows the attenuation provided by the earplug for the respective frequency bands to be determined.
These attenuation values take into account all the subjective aspects of sound perception, but on the other hand, data collection is a time-consuming process and the obtained values are characterized by quite large inter-individual variability.
Therefore, the optimal solution is to measure the attenuation characteristics based on the objective method (without the presence of the subject), the results of which will be in accordance with the results of subjective tests.
The best solution is to use an artificial head (head simulator) that contains full copy of the auricles as well as the ear canal. There are many models of artificial heads with artificial ears currently on the market, however, this equipment does not have a full copy of the ear canal.
One of the most important works on the comparison of research methods to determine the value of the acoustic attenuation of hearing protectors is the paper (Berger, 2005). In this work, the author compared three methods of measuring the effectiveness of hearing protectors, i.e. the real-ear attenuation at threshold -REAT method -ISO 4869-1 standard and two objective methods: microphone in real ear-MIRE, consisting of placing the microphone in near the tympanic membrane, as well as the acoustical test fixtures -ATF, which uses artificial ear systems consisting of a microphone placed in a suitable cover connected with an acoustic coupler. The element of the artificial ear is connected to the amplifier system and the measuring system, which usually provides a comprehensive analysis of signals, mainly in terms of determining sound levels and assessing the spectral structure.
This work focuses on comparing the results obtained by REAT and ATF methods. The level of exposure to noise during an 8-hour day is defined as: where , is the time average sound level determined for the time of noise exposure, T e , and = 8 hours = 28800s. T e is the time in which noise affects the hearing system.
The level of exposure to noise during a 5-day working week is defined by the following equation: where the index i denotes the i-th day in a week, and n = 5denotes the number of working days during the week. The methods of measuring the quantities characterizing noise at workplaces have been presented in detail in (PN-N-01307, 1994). The standard also includes requirements for measuring the equipment, and the mode and frequency of measurements.
The standard presents two methods: -the direct method, which is based on continuous measurement, during the time when the worker is exposed to noise, and on reading the values of determined quantities directly from meters, e.g. noise dosimeters or integrating sound level meters, and -the indirect method, which involves measuring noise in a time shorter than the one being evaluated and applying appropriate mathematical relationships to determine the required quantities.
The When all the technical possibilities to reduce noise at the workplace are exhausted and one of the abovementioned values is still exceeded, the employer is obliged to provide the employees with hearing protection and inform them about the potential risk of hearing damage.

Individual hearing protectors
One way of providing hearing protection at the workplace is through the use of can lead to the effect of excessive protection. This effect is associated with too much noise suppression. This, in turn, may cause the employee to feel acoustic isolation from the surroundings, and to experience reduced communication possibilities and the lack of the ability to hear alarm and warning sounds. As a result, this leads to work discomfort and may lead to the employee's rejection of the hearing protector.
The construction of an individual hearing protector with an acoustic filter is presented in Figure 1. The acoustic filter is mounted in a special sleeve, which is permanently glued to the insert / protector.
To determine the acoustic efficiency (suppression) of individual hearing protectors, subjective and objective methods are used.
The subjective method for measuring the acoustic efficiency of hearing protectors is provided by (PN-EN-ISO-4869-3). This standard describes the subjective method of measuring the acoustic attenuation of hearing protectors at low sound pressure levels (close to the threshold of hearing). The method was developed to obtain attenuation values close to the maximum, which are difficult to achieve in real conditions. The hearing threshold is measured with and without a hearing protector (similar to (PN-EN-ISO-8253-2.)).
The objective method for measuring protectors is given in (PN-EN-ISO-4869-3).
However, this standard applies to ear-muffs.

The design of new acoustic filters
New acoustic filters were designed and manufactured for the purposes of the research. The cross-section of one of these filters is shown in Figure 2. The three filters were designed with different diameters, d, of the inner hole. The filters were labelled with the symbols F1, F2 and F3, however the filter F1 had the smallest inner hole diameter, and F3the largest diameter. The larger the diameter of the inner hole, the smaller the sound reduction.

The determination of the acoustics attenuation individual hearing protectors
Measurements of the acoustic attenuation obtained by individual hearing protectors with designed F1, F2 and F3 acoustic filters, as well as full insert earplugs (without any acoustic filters) were carried out using two methods: objective (chapter 2.3) and subjective (chapter 2.4).

The objective methodmeasurements using a head simulator
The measurements were carried out in an anechoic chamber at the Institute of Acoustics of the Adam Mickiewicz University in Poznan.
The parameters of the anechoic chamber are presented Tab. 2.
The following measuring equipment was used in the measurements: − High-frequency Head and Torso Simulator Brül&Kjaer Type 5128; − Active Loudspeaker QSC type K10; − System Pulse v. 12.6.0.255; − Sound level meter Svantek type SVAN 945A; The measurement system is presented in Figure 3.
The acoustic measurements were made for the following conditions: − without hearing protectors, − with hearing protectors without acoustic filters (the so-called full insert) and with acoustic filters mounted in the insert with the internal holes having different diameters, F1, F2 and F3.
The test signal in the measurements was pink noise -in the frequency range up to 12.5 kHz. The signal was generated by a loudspeaker (Active Loudspeaker QSC type K10). The artificial head (High-frequency Head and Torso Simulator Brül & Kjaer Type 5128) was located at a distance of 3 m, directly opposite the loudspeaker. In addition, a sound level meter was located above the artificial head to calibrate the measuring system.
The pink noise level was 85, 90 and 95 dB. The spectrum of the acoustic signal used in the tests is shown in Figure 4.

The subjective methodmeasurements of the hearing thresholds
The subjective acoustic measurements were carried out in accordance with ISO  Figure 5 and Figure 6 are the audiograms of people participating in the study.
To determine the attenuation value obtained by of hearing protectors, a pink noise signal was used for one-third octave bands: 125, 250, 500, 1000, 2000, 4000 and 8000 Hz.
The attenuation value was defined as the difference (in dB) between the hearing threshold of the test signal with a hearing protector and the hearing threshold determined without a hearing protector.
Measurements were made using an Interacoustics AC40 audiometer. As part of the work, measurements for each subject (for earplugs with different filters) were repeated three times. The level of background noise in the testing room did not exceed the permissible values for background sound pressure levels presented in the standard.

The results of the objective method
As stated earlier, acoustic measurements were made for three scenarios: without ear protectors, with protectors, and with acoustic filters. During the acoustic measurements, the hearing protectors were located on both ears of the artificial head. The measurements were repeated five times for each condition.
The acoustic efficiency (attenuation) of hearing protectors with and without acoustic filters was determined as the difference between the sound levels recorded without a hearing protector and with protectorswith different acoustic filters. Figure 7 shows the attenuation values determined for the protectors with filters F1, F2 and F3, and without a filter.

Statistical analysis
An ANOVA analysis of variance was used to perform the statistical analyses. The dependent variable was the attenuation value of the hearing protector, while the factors were: "frequency", "filter type" and "pink noise level". The overall analysis showed that "frequency" is a factor significantly affecting the attenuation values {F (6) = 4720.91, p <0.001}, similarly to "filter type" {F (3) = 4225.49, p <0.001}. However, the level of stimulation of the used pink noise turned out to be a factor that did not significantly affect the values of the attenuation of the earplugs with acoustic filters and in the case of the full earplug {F (2) = 0.841, p = 0.432}.
The relationship between the average attenuation and the filter type for respective levels of stimulation is given in Figure 8.
The Tukey test showed that at the confidence level p = 0.05, the average values of attenuation for earplugs with filters F1, F2, F3 and full earplugs (FE) differ significantly.

The results of the objective method
The obtained values of acoustic attenuation for individual subjects and the average attenuation values for the full earplugs with F1 -F3 filters are shown in Figure 10 - Figure 13. Figure 14 shows a comparison of the attenuation for individual earplugs with acoustic filters and for the full earplugs. Figure 15 shows the attenuation values averaged over all the frequencies for individual earplugs with F1, F2, F3 filters and for the full earplug.

Statistical analysis
As part of the study, an ANOVA analysis was performed, in which the dependent variable was "attenuation", and the factors were: "frequency", "filter type". The main analysis of variance showed that both factors were statistically significant, at the significance level p ≤ 0.001. This means that both "frequency" {F (6)  This means that for all types of earplugs with filters and the full earplug the value of attenuation changed as a function of frequency, however the average values of attenuation between F1 and F2 were not statistically significant. It can therefore be concluded that the attenuation characteristics of the earplugs with the F1 and F2 filters are comparable. Only the use of the F3 filter significantly changed the attenuation characteristics. Therefore, to attain greater diversity in the attenuation characteristics of the earplugs with filters, a filter with a diameter smaller than the F1 and F2 filters, and larger than the F3 filter, should be used. In addition, a filter with a smaller diameter than F3 should be used to obtain higher attenuation values, especially in the higher frequency band. The above analysis shows that the earmold combined with a properly selected acoustic filter allows to obtain a large diversity of attenuation values in each of the analyzed frequency bands, which cannot be obtained with a full earplug.

A comparison of the results obtained by the objective and subjective methods
The next stage of the analysis consisted in comparing the obtained results of acoustic attenuation for a full earplugs and earplugs with different filters (F1-F3), as obtained by the subjective method and the objective method. Statistical analysis showed that for objective measurements the effect of the level of pink noise stimulation was not statistically significant, therefore this factor was not taken into account in these analyses. The factors were: "method", "frequency" and "filter type".
A general analysis of ANOVA variance showed that "frequency" and "filter type" are factors that significantly affect insert values {F (6)

Discussion
One of the most important papers on measuring the effectiveness of the attenuation caused by of hearing protectors is the publication (Berger, 2005). The author compared three methods of measuring the effectiveness of hearing protectors, i.e.
the real-ear attenuation at threshold (REAT) method based on measuring the hearing threshold of people with hearing protectors, and two objective methods. The first objective method is analogous to that used in this work, based on a measuring system consisting of an artificial ear / artificial head (acoustical test fixtures -ATF).
The second objective method is to measure the sound level in the ear canal after putting on the hearing protector by placing a microphone near the eardrum (microphone in real ear-MIRE). In the conclusion of the work (Berger, 2005), they stated that the REAT method is considered optimal, although it is not without drawbacks, such as those resulting from the need to involve a group of people to perform time-consuming measurements. When the MIRE method is used, bone conduction is not included in the process of acoustic energy transmission.
The optimal solution would be to use the ATF method to measure the attenuation characteristics of hearing protectors, both over-the-ear and in-the-ear, and this has been attempted in this study.
The work (Kozlowski and Mlynski, 2017) focused on determining the attenuation characteristics of ear protectors in the band above 8 kHz, i.e. 10, 12.5 and 16 kHz.
Ear protectors (earmuffs) were used. The subjective (REAT) and objective methods were also used based on the "acoustic text fixture -ATX" -the artificial head of G.R.A.S. 45CB. The authors of this study examined a total of 27 hearing protectors (different companies and types) and obtained large differences in attenuation between individual types of protectors, reaching up to 40 dB at 16 kHz. In addition, they found differences between the attenuation results based on different methods.
For the 10 kHz frequency, the average attenuation values determined by the objective method were higher by 6 dB, while for the 12.5 and 16 kHz frequencies they were greater and these differences on average even amounted to 15 dB. (Chordekar et al., 2016) investigated the effect of factors limiting the attenuation values for earplugs and earmuffs. They assumed that hearing loss, occlusion or bone conduction might be responsible for reducing the attenuation of hearing protectors.
They tested the effectiveness of attenuation provided by hearing protectors using the subjective REAT method. Measurements were made for earmuffs, earplugs and simultaneously used earplugs and earmuffs. The test results showed that the use of ear protectors with the earplugs in place does not change the attenuation efficiency.
They showed that the transmission reduction of vibrations through soft tissues is responsible for influencing a reduction in the attenuation provided by hearing protectors, also in the case of earplugs placed deep in the ear canal. ( Mlynski et al., 2014) studied the effectiveness of attenuation of two types of hearing protectors in the case of impulse noise, which was the impact of a steel hammer at the workplace. In research, they used the objective AFT method. Test results have shown that the highest attenuation efficiency was demonstrated by earplugs.
The article (Bockstael et al., 2008) is an interesting work on the use of hearing protectors in the case of noise from firearms. The authors also measured two types of otoacoustic emissions (distortion product otoacoustic emission -DPOAE, and transiently-evoked otoacoustic emission -TEOAE) in soldiers who used hearing protectors while shooting. The subjects were divided into two groups, one group used passive earplugs, the other one used active earmuffs as hearing protectors. DPOAE and TEOAE measurements were taken before and immediately after shooting for a period of 5 days. The test results showed that, regardless of the type of hearing protectors, no significant changes in OAE levels were found after exposure to noise from firearms. This indicates that the inner ear cochlea is sufficiently protected against noise at too high a sound pressure level.

Conclusions
To sum up, the most important conclusions resulting from this work are as follows:  The results of this study partly confirmed the hypothesis that there is no significant impact of the measurement method on the attenuation characteristics of the earplugs with different types of acoustic filters.
It should be added that the increasingly better systems of artificial ear / artificial head, especially those in which the full projection of the ear canal has been used, gives reason to hope that in the future fully objective measurement of the attenuation provided by any hearing protectors and earplugs will be possible. The research results included in this work are preliminary studies. The plan is to use earplugs with a greater variety of acoustic filters in subsequent research and to significantly increase the group of people surveyed with the subjective method.