Effects of Tonal Noise on Workers’ Annoyance and Performance

Objective: Numerous references indicate that the subjectively assessed tonal noise annoyance is higher than that of broadband noise. There are no criteria for the impact of tonal noise in assessing the occupational environment for both indoor areas and workplaces. Materials and Methods: The study participants included 50 people who met the audiometric qualification criteria. The research method employed both a questionnaire survey and computer psychological tests checking work performance, attention level, and memory. Four types of generated test signals were developed (filtered noise − A and three signals with tonal components 125, 1600, and 8000 Hz − B, C, and D) at the same sound level A of 55 dB. Test signals C and D were assessed as causing the greatest annoyance and as the loudest. Results: The results of some tests and the assessment of annoyance and of the volume of test signals containing medium and high frequency tonal components were correlated with the participants’ noise sensitivity, determined on the basis of a questionnaire. Although there are no statistically significant differences, it was observed in most cases for signals with C (1600 Hz) and D (8000 Hz) tonal components that the results (mean values or median values) of psychological tests deteriorated with respect to a noise signal without tonal components (A) − a smaller number of calculations, a smaller number of correct responses, more errors made. Conclusion: These results, combined with those of the questionnaire survey, justify the introduction of the tonality annoyance criterion for workstations where, among other things, focusing one’s attention is required.

Keywords: Annoyance, noise, psychological tests, tonality, workstation

How to cite this article:
Radosz J. Effects of Tonal Noise on Workers’ Annoyance and Performance. Noise Health 2021;23:117-27
Introduction

Tonal noise can be defined as noise with tonal frequency components in its spectrum. In addition to damaging the hearing organ, as a stressor, noise can contribute to the development of adverse health effects (e.g., hypertension, coronary heart disease), can impair cognitive performance, cause distraction, hinder work, and reduce staff performance.[1],[2],[3] According to the ISO/TS 15666[4] definition, annoyance being a consequence of noise is an individual, adverse reaction of a person causing dissatisfaction, anxiety, irritation, or disturbance. The World Health Organization associates noise annoyance with an adverse effect on health and defines it as the experiencing of many different reactions such as anger, disappointment, dissatisfaction, withdrawal, helplessness, depression, fears, distraction, or fatigue.[5] Noise annoyance can also be determined as a psychological concept associated with disturbance, aggravation, dissatisfaction, concern, bother, displeasure, harassment, irritation, nuisance, vexation, exasperation, discomfort, uneasiness, distress, and hate.[6] Loudness seems to be the most significant noise characteristic associated with the perception of annoyance. Loudness is commonly evaluated by using the following parameters: A-weighted equivalent sound pressure level (LAeq), loudness levels as per ISO 532[7] and ANSI 3.4-2007,[8] and even the perceived noise level.[9] However, many studies indicate that only a fraction of the perception of annoyance can be predicted by loudness. Brocolini et al.[10] suggested that only approximately 30% of annoyance results from loudness due to other sound characteristics and nonacoustic factors.[6],[11] Relevant literature data clearly indicate that the tonality of noise is one of the most important remaining noise characteristics that should be taken into account when evaluating annoyance.[12]

Technical equipment components of a building, such as heating, ventilation or air-conditioning, are increasingly energy efficient; however, less attention is given to their acoustic quality.[12],[13] A large part of this equipment emits clear tonal sounds as a result of rotation of such parts as fans and pumps. Tonal sound may be also emitted by office equipment − computers, printers, and telecommunications equipment.[14]

A number of studies on the link between tonal noise and man felt annoyance have been published; however, research findings have been inconsistent and limited. Some studies did not find any statistically significant differences in task performance between broadband and tonal noises,[15],[16],[17] although Lee et al.[18] found that there was a trend of decreasing accuracy with increasing tone strengths. Landström et al.[19],[20] studied noise levels and annoyance at real workplaces. The observed correlation between a noise level and annoyance was weak, but the perceived annoyance increased significantly when tones were present in the noise spectrum. These studies suggest that the tonality indicators should be taken into account in an assessment of the noise generated by the technical equipment in the building, but none of the existing tonal indicators has been widely used so far.

Also, there is no information on the tonality indicators with regard to the permissible noise levels at workstations.[21] Recent literature reports[18],[22] suggest that there is a need to develop criteria for tonal noise annoyance; however, the results of recent research are not sufficient to establish such criteria. Therefore, the purpose of this study was to determine the impact of signals with different tonal characteristics on the results of standardized psychological tests.

Materials and methods

The sample selection in the laboratory tests employed a nonrandom method consisting in establishing in advance the characteristics to be met by individuals in the sample (age below 50 and audiometric qualification). The structure of the sample in this case was established arbitrarily. Fifty people (25 men and 25 women) took part in the tests. The mean age was 30.4 (range 20–50). At the beginning, each study participant underwent tonal audiometry. The test qualification threshold was adopted as 20 dB hearing loss (HL) in the considered frequency range.

Before the tests started, every subject was asked to fill in a noise sensitivity questionnaire. An abbreviated questionnaire (NoiSeQ) developed by Schütte was used for this purpose.[23] Each subject was acquainted with the test procedures, received proper instruction on how to perform the tasks, and had to sign an informed consent form to participate in the tests. Anonymity procedures were applied to protect the privacy of human subjects while collecting, analyzing, and reporting data

The main part of the test consisted of four parts broken down by the test signal generated. In each part, the subjects performed specific tasks on a computer (psychological tests), and then they assessed the test signals by using the questionnaire that included the assessment of the annoyance (verbal scale as per ISO 15666[4]) and loudness of the signal generated (10 points scale).

The Vienna Test System was used in the study and three types of psychological tests were selected to assess the performance of the participants while exposed to the test signals generated:

Attention and concentration test (COG), S4 form

Attention continuity test (DAUF), S1 form

Work performance test (ALS), S7 form

The test scenario included the following four types of test signals:

Signal A − filtered pink noise (LAeq = 55 dB)Signal B − filtered pink noise with the tone at a frequency of 125 Hz (LAeq = 55 dB, tone-to-noise ratio [TNR] = 19.1, prominence ratio [PR] = 20.1)Signal C − filtered pink noise with the tone at a frequency of 1600 Hz (LAeq = 55 dB, TNR = 10.9, PR = 12.5)Signal D − filtered pink noise with the tone at a frequency of 8000 Hz (LAeq = 55 dB, TNR = 11.9, PR = 12.5)

The first assumption for the measurement signals developed was to ensure that each of them had the equivalent sound pressure level A of 55 dB (the criterion for workstation noise annoyance[24],[25]). The second assumption was that for signals with tonal components the TNR and PR parameters should be 3 dB above the limits of criterion curves, which would prove that the tonal components in the spectrum are significant.

The signal without any tonal components (filtered pink noise) had its frequency corrected in such a manner as to ensure that the spectrum shape corresponded to noise criterion (NC) curves.[26] Following this correction, the NC value calculated for signal A was 46 (for LAeq = 55 dB). The remaining signals were generated by adding tonal components to signal A and adjusting the proportions accordingly to the assumed overall signal loudness.

During exposure to each of the test signals, the subjects performed three psychological tests (COG, DAUF, and ALS) and then they assessed a given signal in the questionnaire.

The sequence of signal presentation was based on the Latin square design to exclude the effect of the signals sequence on the assessment results. The time of a full test for one participant was approximately 2 hours.

The test stand consisted of

Interacoustics AD629 diagnostic audiometer,

Tabor Electronics WW5062 signal generator,

Behringer DEQ2496 Ultra-Curve Pro digital equalizer,

PC with Schuffried software − Vienna Testing System,

a set of three active M-Audio BX5 monitor speakers (bandwidth 56 Hz − 22 kHz), and

Svantek SVAN 979 SOUND level meter/analyzer for monitoring the generated signals.

The test-retest method was used to assess the reliability of the research method. The results of the test method reliability evaluation have been presented in earlier publications.[27]

The following statistical tests were used to analyze the test results: Shapiro-Wilk test to evaluate the normality of the variable distribution, Mauchly test to check the spherical assumption, analysis of variance (ANOVA) for dependent groups (interval-scaled variables), unifactorial multivariate analysis of variance (MANOVA) (interval-scaled variables if the spherical assumption is not met), and ANOVA Friedman test (ordinal-scaled variables and variables if the distribution normality assumption is not met). Statistica 13 and PQStat 1.6.8 software was used in the statistical analysis. A significance level of 0.05 was adopted in the analysis.

Results

First, the results of the questionnaire survey are discussed. The results of the subjective assessment of the annoyance of the signals under consideration are presented in [Figure 1]. The signals examined were perceived as lying within the range from slightly to extremely annoying. On average, signals A and B were rated as moderately annoying, while signals C and D as very annoying.

Figure 1 Results of subjective assessment of signal annoyance (1–not annoying at all; 2–slightly annoying; 3–moderately annoying; 4–very annoying; 5–extremely annoying)

Click here to view

The results of the subjective assessment of the sound loudness are presented in [Figure 2]. A considerable span between individual ratings was observed (from 0 to 10 on the evaluation scale). On the other hand, average score values indicate moderate volume of the tested signals (score of 4 for signals A and B and score 5 for signals C and D).

Figure 2 Results of subjective assessment of signal loudness under tested conditions

Click here to view

Next, the results of the DAUF, COG, and ALS psychological tests were discussed. For the DAUF test, the results of the sum of correctly and incorrectly performed tasks and the average time of task performance were discussed. [Figure 3] presents the results of the sum of correctly performed tasks in the DAUF test. The largest span between results was observed for signal A (from 106 to 120 tasks). Average values for individual signals ranged from 114 to 116 tasks.

Figure 3 Results of the DAUF attention continuity test − the sum of properly performed tasks

Click here to view

[Figure 4] shows the average time needed for proper performance of a task in the DAUF test. The average times needed for proper task performance ranged from approximately 0.6 s to approximately 1.3 seconds. Average values lay within the range from 0.8 to 0.9 second for individual signals. [Figure 5] presents the results for the sum of incorrectly performed tasks in the DAUF test. The largest span between results was observed for signal A (from 1 to 6 tasks). Average values for individual signals ranged from 2 to 4 tasks.

Figure 4 Results of the DAUF attention continuity test − mean time needed for proper test performance

Click here to view

Figure 5 Results of the DAUF attention continuity test − the sum of improperly performed tasks

Click here to view

For the COG concentration and attention test, the results of the sum of properly and improperly performed tasks and the mean time for correctly rejected figures were discussed.

[Figure 3],[Figure 4],[Figure 5],[Figure 6] present the results for the sum of correctly performed tasks in the COG test. Mean values for individual signals ranged from 57 to 69 tasks.

Figure 6 Results of the COG attention and concentration test − the sum of properly performed tasks

Click here to view

The results of the ALS performance test are shown in [Figure 7]. The range of the sum of the tasks performed in the ALS test for all the signals lay between 314 and 437. Average sums ranged from 353 to 395 for individual signals.

Figure 7 Results of ALS performance test − sum of performed operations

Click here to view

The results of the Shapiro-Wilk tests indicate that only for the ALS test (number of performed calculations) and the DAUF test (average time of correct reactions) for signals B and C are there no grounds for rejecting the hypothesis of distribution normality for the tested variables. The results also indicate that for the ALS test (number of performed calculations) the assumption of the variance sphericity is not met. Therefore, the ANOVA multivariate analysis, which does not assume sphericity, was used for this test (F = 5.3, P = 0.03).

The results of the MANOVA showed statistically significant differences between mean values for different signals in the ALS test (the number of performed calculations). The results of the post hoc analysis [Fisher Least Significant Difference (LSD)] showed statistical differences between signals A and B (on average by approximately 19) and B and D (on average by approximately 15). In the case of other psychological tests, the repeated measures analysis of variance for Friedman ranks (ANOVA Friedman) did not demonstrate a statistically significant difference between the test results depending on the test signals [see [Table 1]].

The results of the ANOVA Friedman test showed statistically significant differences between individual signals for all questionnaire tests. The Dunn-Bonferroni post hoc analysis revealed the following:

Statistically significant differences between signals A-D, B-C and B-D when assessing signal annoyanceStatistically significant differences between signals in an assessment of how demanding were the tasks performed (Dunn-Bonferroni post hoc analysis did not show any statistically significant differences between signals; however, the values at the limit of the confidence level indicate differences between signals A-D and B-DStatistically significant differences between signals A-D, B-C and B-D when assessing the signal loudnessStatistically significant differences between signals A-C, A-D, B-C and B-D when assessing difficulties in the task performance.

Owing to the lack of normality of dependent variable distributions and the lack of statistically significant diversity for most psychological tests, an analysis of outlying observations was performed. A three-sigma rule was used to detect outlying observations; it was reiterated until all outlying observations were replaced by no data. Outlying observations were not detected for ALS tests. Then, a repeated measure analysis was performed for the Skillings-Mack rank, which can be used in the absence of data. Post hoc analysis is not performed for this type of test. The results are presented in [Table 2] and they do not indicate a statistically significant difference between the signals for the psychological tests considered in this analysis.

A general linear model factorial analysis of variance of was used to analyze the impact of several factors (sensitivity to noise, age, sex) on the level of psychological test performance. It makes it possible to use multiple factors to identify the compared groups. The study may also include variables that constitute an interaction of the indicated factors. Factorial ANOVA requires factors to be divided into different categories (i.e., independent populations). Independent variables are categorized as follows:

Sex − male (M), female (K)Noise sensitivity (based on noise sensitivity questionnaire) − low (global index ≥3), moderate (2 ≤ global index < 3), high (global index <2)Age − 20-30, 30-40, 40-50

The analysis was performed by using the third type of the sum of squares and the encoding of the effects. Given the data collected, only sex/sensitivity and sex/age interactions were included in the analysis.

The results for the ALS test are shown in [Figure 8] and [Figure 9]. For signal A, statistically significant differences were observed in the results of the ALS test (the number of performed calculations) between participants of different noise sensitivity (variability of approximately 18%), between participants of different ages (variability of approximately 43%), and between participants of different sexes depending on the noise sensitivity (variability approximately 18%). The test result was not influenced by sex (variability below 5%) or sex depending on age (below 1%).

Figure 8 ALS test result (number of calculations performed) for signals A to D depending on participants’ noise sensitivity

Click here to view

Figure 9 ALS test result (number of calculations performed) for signals A-D depending on the participants’ noise sensitivity

Click here to view

The results of the post hoc analysis showed statistically significant differences for the following:

ALS test (the number of calculations performed) for signal A between individuals aged from 20 to 30 years and individuals aged from 30 and 40 years (on average by approximately 90) and between individuals aged from 20 to 30 years and individuals aged from 30 to 40 years (on average by approximately 105)ALS test (the number of calculations performed) for signal A between high-noise-sensitive women and low-noise-sensitive men (on average by approximately 126)

For signal B, statistically significant differences were observed in the results of the ALS test (the number of performed calculations) between participants of different noise sensitivity (variability of approximately 18%), between participants of different ages (variability of approximately 41%), and between participants of different sexes depending on the noise sensitivity (variability of approximately 20%). The test result was not influenced by sex (variability of approximately 5%) or sex depending on age (approximately 6%).

The results of the post hoc analysis showed statistically significant differences for the following:

ALS test (the number of calculations performed) for signal B between individuals aged from 20 to 30 years and individuals aged from 30 to 40 years (on average by approximately 93) and between individuals aged from 20 to 30 years and individuals aged 40 to 50 years (on average by approximately 100)ALS test (number of calculations performed) for signal B between high-noise-sensitive women and low-noise-sensitive men (on average by approximately 144), between high-noise-sensitive women and high-noise-sensitive men (on average by approximately 136), and between low-noise-sensitivity men and moderate-noise-sensitivity men (average by approximately 70)

For signal C, statistically significant differences were observed in the ALS test results (number of calculations performed) between individuals of different noise sensitivity (variability of approximately 23%), between individuals of different ages (variability of approximately 32%), and between individuals of different sexes depending on the noise sensitivity (variability approximately 18%). The test result was not influenced by sex (variability of approximately 3%) and sex depending on age (approximately 6%). The results of the post hoc analysis showed statistically significant differences for the following:

ALS test (number of calculations performed) for signal C between low-noise-sensitivity individuals and moderate-noise-sensitivity individuals (on average by approximately 70)ALS test (number of calculations performed) for signal C between individuals aged from 20 to 30 years and individuals aged from 30 to 40 years (on average by approximately 68) and between individuals aged from 20 to 30 years and individuals aged 40 to 50 years (on average by approximately 81)ALS test (number of calculations performed) for signal C between high-noise-sensitive women and low-noise-sensitive men (on average by approximately 132) and between low-noise-sensitivity men and moderate-noise-sensitivity men (average by approximately 90)

For signal D, statistically significant differences were observed in the results of the ALS test (number of calculations performed) between individuals of different noise sensitivity (variability of approximately 27%), between individuals of different ages (variability of approximately 34%), and between individuals of different sexes depending on the noise sensitivity (variability of approximately 17%). The test result was not influenced by sex (variability of approximately 3%) and sex depending on age (approximately 6%). The results of the post hoc analysis showed statistically significant differences for the following:

ALS test (number of calculations performed) for signal D between low-noise-sensitivity individuals and moderate-noise-sensitivity individuals (on average by approximately 77) and between low-noise-sensitive individuals and high-noise-sensitive individuals (on average by approximately 89)ALS test (number of calculations performed) for signal D between individuals aged from 20 to 30 years and individuals aged from 30 to 40 years (on average by approximately 78) and between individuals aged from 20 to 30 years and individuals aged from 40 to 50 years (on average by approximately 73)ALS test (number of calculations performed) for signal D between high-noise-sensitive women and low-noise-sensitive women (on average by approximately 146), between high-noise-sensitive women and low-noise-sensitive men (on average by approximately 151), and between low-noise-sensitivity men and moderate-noise-sensitivity men (average by approximately 83)

For the COG test (number of correct responses), statistically significant differences were observed in the test results between individuals of different ages for all signal types (variability of approximately 16% for signal A, approximately 29% for signal B, approximately 18% for signal C, and approximately 16% for signal D) [Figure 10].

Figure 10 COG test result (number of correct responses) for signals A-D depending on the age of the tested persons

Click here to view

The results of the post hoc analysis showed statistically significant differences for the COG test (number of correct responses) for signals A to D between individuals aged from 20 to 30 years and individuals aged from 40 to 50 years (on average by approximately 8 for signal A, by approximately 8 for signal B, approximately 9 for signal C, and by approximately 8 for signal D).

For the COG test (number of wrong responses), statistically significant differences were observed in the test results between individuals of different ages depending on the noise sensitivity for the three types of signal (variability of approximately 15% for signal A, approximately 15% for signal B, and approximately 17% for signal D). As the noise sensitivity increases, the number of wrong responses increased for women and decreased for men for all types of signals.

In the other tests, no statistically significant impact of the factors under consideration on the results for all types of signal was observed:

ALS test (error percentage rate) − maximum variability below 6%DAUF test (number of correct responses) − maximum variability below 6%DAUF test (number of wrong responses) − maximum variability below 6%DAUF test (mean correct response time) − maximum variation below 6% Discussion

The results of this study showed that the ALS work performance test, which examined the ability to concentrate on performing calculation tasks with a time load, turned out to be a particularly interesting psychological test used to assess the impact of tonal noise on the tasks performed by an employee. Among all psychological tests, only in this test were there no grounds for rejecting the distribution normality hypothesis for the tested variables for all types of test signals.Also, only for this test did the applied factor model, in which the impact of independent variables (noise sensitivity, age, sex) on the test result was examined, show statistical significance. The majority of previous studies have mainly focused on the effects of noise sensitivity on annoyance ratings for outdoor noises such as environmental noise (road traffic noise,[28] aircraft noise,[29],[30] etc.). According to these reports, noise sensitivity can be independent predictor of annoyance and explains much of the variance between noise exposure and individual annoyance responses.[29] The results presented in the paper support the previous findings that noise sensitivity influences reactions to noise exposure (both tonal sounds and sound without tonal components).

The test signals were developed under an assumption that the A-weighted equivalent sound pressure level for each of them was 55 dB (the criterion of noise annoyance at the workstation in administrative rooms, design offices, rooms for theoretical work, data processing and other for similar purposes in accordance with Polish[24] and German regulations[25]). The tonal component signals differed in the tone frequency (125 Hz; 1600 Hz; 8000 Hz) and the tonality parameters for the prominence criterion curves given in ANSI/ASA S12.10-2010 or ECMA-74. Although signal B with the 125 Hz tonal component had higher values for TNR and PR than the other signals C (1600 Hz) and D (8000 Hz), it caused the least annoyance as assessed by the tested individuals. It was also rated as the least loud. In the case of the assessment of how demanding and how difficult the task performed was under the test conditions, it was comparable to the noise signal (signal A). Test signals with 1600 Hz (C) and 8000 Hz (D) tonal components were assessed as causing the greatest annoyance and as the loudest. It was similar in the assessment of how demanding and difficult the task to be performed was under the examined conditions. The assessment of annoyance and of the loudness of test signals with medium and high frequency tonal components were correlated with the tested individuals’ sensitivity to noise, determined on the basis of a questionnaire. The results of the study did not allow for unambiguous determination of the impact of age and sex on the perceived annoyance of tonal noise.

Conclusions

In most cases for tested signals with the medium and high frequency tonal components, the results of psychological tests deteriorated with respect to a signal without tonal components − a smaller number of calculations, a smaller number of correct responses, and more errors made. These results, combined with those of the questionnaire survey, justify the introduction of the tonality annoyance criterion for workstations where, among other things, focusing one’s attention is required.

One possibility of establishing the tonal noise annoyance criterion is to introduce adjustments to the measured sound pressure levels. In German regulations, an adjustment depending on the subjective feeling of the employee concerning noise annoyance is added to the measured A-weighted sound pressure level. However, those provisions are fairly general and do not specify the method of subjective assessment. The adjustment for tonal noise annoyance may depend on the resulting global value for the noise sensitivity questionnaire, which can be regarded as an additional tool for workplace assessment.

Financial support and sponsorship

This paper has been based on the results of a research task carried out within the scope of the fourth stage of the National Programme “Improvement of safety and working conditions” partly supported in 2017-2019 — within the scope of state services — by the Ministry of Family, Labour and Social Policy. The Central Institute for Labour Protection − National Research Institute is the Programme’s main co-ordinator.

Conflicts of interest

There are no conflicts of interest.

References

1.Saeki T, Fujii T, Yamaguchi S, Harima S. Effects of acoustical noise on annoyance, performance and fatigue during mental memory task. Appl Acoust 2004;65:913-21.

2.Belojevic G. Noise and performance: research in Central, Eastern and South-Eastern Europe and newly independent states. Noise Health 2013;15:2-5.

[PUBMED] [Full text] 3.Basner M, Babisch W, Davis A et al. Auditory and non-auditory effects of noise on health. Lancet 2014;383:1325-32.

4.International Organization for Standardization, ISO/TS 15666:2003 Acoustics − Assessment of Noise Annoyance by Means of Social and Socio-Acoustic Surveys, 2003.

5.WHO. Burden of disease from environmental noise − quantification of healthy life years lost in Europe; n.d. Available from https://www.who.int/quantifying_ehimpacts/publications/e94888/en/. Accessed November 18, 2019.

6.Guski R. Personal and social variables as co-determinants of noise annoyance. Noise Health 1999;1:45-56.

[PUBMED] [Full text] 7.International Organization for Standardization. ISO 532:1975 Acoustics — Method for calculating loudness level, 1975.

8.American National Standards Institute. ANSI S3.4-2007 Procedure for the Computation of Loudness of Steady Sounds (Includes Loudness Program), 2007.

9.Kryter KD. The meaning and measurement of perceived noise level. Noise Control 1960;6:12-27.

10.Brocolini L, Lavandier C, Marquis-Favre C, Quoy M, Lavandier M. Prediction and explanation of sound quality indicators by multiple linear regressions and artificial neural networks, Proceedings of the Acoustics 2012 Nantes Conference, 2012, p. 2121-6.

11.van Kamp I, Davies H. Noise and health in vulnerable groups: a review. Noise Health 2013;15:153-9.

12.Lee KH, Davies P, Surprenant AM. Quantification of the Tonal Prominence of Complex Tones in Machinery Noise, Proceedings of Noise-Con 04, 2004;114:927-33.

13.Pleban D. Definition and measure of the sound quality of the machine. Arch Acoust 2014;39:17-23.

14.Sottek R. Calculating tonality of IT product sounds using a psychoacoustically-based model, Proceedings of Internoise 2015 – 44rd International Congress on Noise Control Engineering, 2015;1:3712-23.

15.Ryherd EE, Wang LM. Implications of human performance and perception under tonal noise conditions on indoor noise criteria. J Acoust Soc Am 2008;124:218-26.

16.Holmberg K, Landstrom U, Kjellberg A. Effects of ventilation noise due to frequency characteristic and sound level. J Low Freq Noise Vib Act Control 1993;12:115-22.

17.Smagowska B, Pawlaczyk-Łuszczyńska M. Effects of ultrasonic noise on the human body-a bibliographic review. Int J Occup Saf Ergon 2013;19:195-202.

18.Lee J, Francis J, Wang L. How tonality and loudness of noise relate to annoyance and task performance. Noise Control Eng J 2017;65:71-82.

19.Landström U, Kjellberg A, Byström M. Acceptable levels of sounds with different spectral characteristics during the performance of a simple and a complex non-auditory task. J Sound Vib 1993;160:533-42.

20.Landström U, Åkerlund E, Kjellberg A, Tesarz M. Exposure levels, tonal components, and noise annoyance in working environments. Environ Int 1995;21:265-75.

21.European Agency for Safety and Health at Work. 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), 2003.

22.Lee J. The effects of tones in noise on human annoyance and performance. Archit Eng - Diss Stud Res 2016;39.

23.Schütte M, Marks A, Wenning E, Griefahn B. The development of the noise sensitivity questionnaire. Noise Health 2007;9:15-24.

24.Polish Committee for Standardization. PN-N-01307:1994 Noise − Noise limit values in the work environment − Requirements for measurement, 1994.

25.German Federal Immission Control Act. Technical Instructions on Noise Abatement (TA Lärm), 1998.

26.American National Standards Institute. ANSI/ASA S12.2-2008 American National Standard Criteria for Evaluating Room Noise, 2008.

27.Radosz J. Pilot study on the influence of tonal noise annoyance on work performance. Proceedings of 25th International Congress on Sound and Vibration, 2018;4:2520-7.

28.Stansfeld SA, Sharp DS, Gallacher J, Babisch W. Road traffic noise, noise sensitivity and psychological disorder. Psychol Med 1993;23:977-85.

29.Van Kamp I, Job RFS, Hatfield J, Haines M, Stellato RK, Stansfeld SA. The role of noise sensitivity in the noise-response relation: a comparison of three international airport studies. J Acoust Soc Am 2004;116:3471-9.