The contribution of personal audio system use and commuting by bus on daily noise dose.

BACKGROUND: For many young people, exposure to music from personal audio system use may represent a significant component of daily noise dose. Moreover, there is increasing concern for the hearing of those who listen at high volumes. The purpose of this study was to determine the noise levels experienced on commuter buses, and to investigate how these impact on the volume-setting behavior of young adult personal audio system users.
METHODS: A questionnaire was used to probe transport use, personal audio system-listening behaviors and the extent of understanding about noise-induced hearing loss. The influence of bus noise on volume-setting behavior was determined by measuring, in a lab setting, the sound-level preferences of participants when listening to their favorite song, a generic song, or a podcast in the absence and presence of various levels of bus noise, simulated using output-adjusted recordings made of bus noise. Statistical analysis was conducted using analysis of variance.
RESULTS: While the bus noise itself was below 85 dB Leq, as the sound level of the buses increased, so did the percentage of commuters who were found to exceed the equivalent of 8 hours of exposure at 85 dB Leq. IMPLICATIONS: Investment in buses with lower noise levels or the use of noise-canceling or noise-occluding headphones would help to reduce the likelihood of noise-induced hearing loss for bus commuters.

Key Messages: Many young commuters would receive the equivalent of 85 dB Leq, 8 hours or more from an hour of travel on a bus when using a personal audio system. This shows that public health initiatives are required to reduce the risk of noise-induced hearing loss in public transport users.Text

INTRODUCTION

Noise-induced hearing loss (NIHL) is recognized as one of the leading causes of hearing loss, and occurs due to repeated exposure to high levels of sound.[1] As hearing loss due to ageing is also important, NIHL accounts for approximately 16% of all disabling loss in adults, and has long been recognized as a costly health issue.[1]

The New Zealand Health and Safety at Work Act[2] includes the requirement for noise assessments and the provision of hearing protection by employers. The standard currently states that an employee can be exposed for up to 85 dBA Leq over an 8-hour period and a peak of 140 dB without exceeding the maximum daily allowable noise dose and being required to use hearing protection. If an average level of 85 dBA Leq is exceeded, then protective measures such as a redesign of equipment, the introduction of acoustic shielding, or the use of hearing protection equipment are to be enforced to attenuate the sound being received by the individual to a level that is unlikely to result in a significant level of hearing loss.

As well as NIHL, the observation of noise-induced synaptopathy and neuropathy in young animals[3] has given rise to concerns about the possibility of an association in humans between noise exposure and so-called hidden hearing loss, a difficulty hearing in a background of noise, despite having normal hearing in quiet, as identified by pure-tone audiometry. The link is consistent with the observation[3] of a loss of high-threshold auditory nerve fibers after noise exposure because of the likelihood that these fibers mediate the signals presented in noise where lower-threshold fibers are saturated. In research conducted in small mammals, there is evidence of cochlear synaptopthy and neuropathy after a single short (2-hour) exposure to sounds at levels of 100 dB, well within the output range of personal audio systems.[3] Furthermore, the lost fibers appear to be those with low spontaneous firing rates,[4] possibly implicating this type of injury in “hidden” hearing losses that are experienced in higher background noise levels but not in quiet conditions. Concern has therefore arisen about this process occurring in humans, where, for ethical reasons, research has been carried out by selecting groups with or without “natural” exposure to high-level sound. Some findings have suggested that the ratio of cochlear activity (as measured by the summating potential) to auditory nerve activity (as measured by the compound action potential) was higher in participants who had not been exposed to much high-level sound compared to those who had.[5] This may imply a loss of auditory nerve fibers, though the effect was driven partly by an increased summating potential, which is not predicted by the theory. A review of the 17 studies published about human research attempting to investigate the phenomenon to date found that only five of these presented results consistent with the synaptopthy/neuropathy effect observed in smaller mammals.[6]

As much exposure to noise occurs in the workplace, for many people, most of the daily noise dose is accumulated as a result of leisure-time activities. In particular, the use of personal audio systems has become an activity of concern. When used in an environment where the background noise levels are high, for example, during commuting or when at a gym, personal audio system volume settings may be increased, leading to a greater risk of high noise doses being experienced. A nationwide study conducted in Australia found that on an average, 62% of those who use personal audio systems use it while commuting, and 16% of listening time was spent while commuting. Of all age groups, those 15 to 19 years old spent the most amount of time listening to their personal listening devices (88 hours per month on an average), followed by those 20 to 29 years old (72 hours per month).[7] Moreover, it has been found that an increasing level of background noise is associated with an increase in the volume setting chosen by listeners, further contributing to their noise dose.[8]

The aim of the present study is to determine the extent to which the sounds produced by buses impact on the preferred listening levels of bus commuters, and hence the impact of bus noise on daily commuting noise dose. Though numerous studies have looked at the external environment and the impact of road traffic noise, few studies have focused specifically on the on-board noise levels experienced inside vehicles. One study carried out in Brazil documenting noise levels on city buses showed that when the levels were not sufficiently high in themselves to produce significant NIHL, the levels were generally over 65 dBA, a level considered high from the point of view of irritation.[9] However, a separate study carried out in a major city of Brazil recorded bus noise levels ranging from 78 to 84 dBA (with a mean of 81 ± 1 dBA),[10] similar to the levels found in a more recent study in large city in Africa of 78 dBA when the windows on the bus were closed and 83 dBA when the windows on the bus were open.[11] In a study conducted in Dhaka in Bangladesh, it was found that levels aboard city buses exceeded 85 dBA 15% of the time.[12] However, the extent to which bus engine noise impacts on the preferred listening levels of personal audio system users who are passengers on such buses is yet to be studied. This study addresses this research gap, as well as investigating whether hearing loss and/or difficulty identifying speech in noise are associated with personal audio-listening system-mediated exposures.

METHODS

Overview

The study consisted of sound measurements made aboard city buses, the administration of a questionnaire probing both personal audio system use and the travel behaviors of young adult participants traveling aboard buses, and a laboratory-based study carried out on volunteers consisting of young adults assessing their personal audio system volume setting behaviors in the presence of varying levels of bus noise administered as background noise. Ethical approval was obtained from the University of Auckland Human Participants Ethics Committee (UAPEC) Reference 2009/263.

Measurement location

Measurements were made aboard city commuting buses in Auckland, New Zealand, using a Bruel and Kjaer 4443 noise dosimeter. Firstly, measurements were taken to determine the extent to which seating location on a bus impacts on noise levels. Five buses were chosen for this purpose, and a series of 5-minute recordings were obtained while sitting at the front, middle, and back of the bus. Recordings were obtained during both the morning and evening commuting periods. On the basis of these measurements, it was found that sound levels did not vary significantly between seating locations on the bus (F = 2.18, P = 0.16). However, for consistency, subsequent data collection was made exclusively from the middle of the bus.

Bus route selection

In total, five representative bus routes were chosen across five areas of the city (Central, North, East, West, and South). For each area, at least two of the routes were journeys traveling at peak traffic times. For each commute, both an A-weighted Leq and a C-weighted peak (LCpeak) were established over a period of approximately 20 minutes. Simultaneously, an M-Audio Microtrak II digital recorder was used to obtain audio recordings of the journey.

Questionnaire

While on board the buses, a short anonymous questionnaire was administered to passengers observed to be using personal audio systems. In total, 38 participants completed the questionnaire, handed to them upon verbal agreement that they would be willing to participate. This questionnaire was designed to establish how long on average people spend commuting by public transport, how much of this time was spent listening to their personal audio systems, as well as what people typically listened to (music or podcasts). It was also used to gain a sense of commuters’ knowledge of, and attitudes toward, NIHL. It was purposely made short, so that people would be able to quickly fill it in and not have to worry about missing their stop, thus consisted of only 10 short questions.

Laboratory-based participation

For the lab-based component, 16 people between the ages of 16 and 25 years were selected to participate (from a total of 18 volunteers). The two exclusions were based on evidence of hearing loss. This age group was chosen due to their increased likelihood of using personal audio systems, and because of their tendency to listen to music at higher volume settings compared with older adults.[13]

The hearing thresholds of participants were determined using standard clinical procedures, including otoscopy, immitance testing, and then pure-tone audiometric threshold testing in a sound-attenuating chamber. Additionally, the QuickSIN test[14] was performed to identify any impaired speech-in-noise ratios (SNRs), a measure of the level of speech required relative to background, in order for speech to be comprehensible. This was carried out to test whether there was a relationship between a poor SNR rating and personal audio system volume setting.

Laboratory setup

The final stage of testing was carried out in a laboratory setting with multiple speakers arranged in a semicircular manner, with the participant sitting in a chair in the middle. From the 20-minute recordings made while on-board public transport, 3-minute segments, devoid of any intelligible speech, were chosen. These three levels represented “loud” (81.2 dBA Leq), “medium” (77.1 dBA Leq), and “quiet” (71.7 dBA Leq) in relation to the average intensity level of the bus sound over the 3-minute period, with the “quiet” and “medium” representative of the low and high end of the noise levels observed on buses in Auckland, and “high” representative of the noise levels of noisy city buses reported on in the international literature. The volumes of the speakers were set to ensure that the participant was exposed to the equivalent noise levels that would be experienced on an actual commute.

For the data collection, a commonly used personal audio system was used, one that did not have a volume display, so participant behavior was not able to be influenced by the displayed volume setting. To measure personal listening volumes, a sound level meter (Bruel and Kjaer 2255) was attached to an artificial ear, and a 2cc insert-type coupler was used with adhesive to hold the earbud in place and in such a way that the earbud would not be completely sealed off to better represent how an earbud typically sits within an ear canal. To correct the sound levels measured in the participants’ ear canals to the equivalent sound pressure of an externally measured sound in a free field, a correction of 8 dB was applied to the Leq measurements (NZS 1269.1:2005).

Testing

Each participant was played two songs and one segment of a podcast for each of the commuter recordings. One of the songs was labeled “generic” to allow direct comparison of sound levels but which might have appealed more to some participants than others. The song chosen was Open Happiness by Cee-Lo Green, Bredon Urie, Patrick Stump, Janelle Monae, and Travis McRoy, and the other (the “favorite”) was chosen by the participant and thus varied between participants. The podcast segment was from a radio talk show and consisted entirely of speech.

The participant first listened to the three songs/podcast in a setting with no significant amount of background noise. During each, the participant was asked to set their personal audio system to their preferred volume setting. After each song/podcast, the output from the personal audio system was measured using the artificial ear and the sound level meter using the volume setting chosen by the participant. The volume setting was then reset, and the trial repeated for the three levels of bus noise (“quiet,” “medium,” and “loud”) for each of the three song/podcast recordings, administered in pseudo-random order.

Statistical analysis

The main analysis was conducted using a two-way (four by three) repeated-measures analysis of variance. The dependent variable was the output level of the personal audio system and the independent variables were the background noise (four levels: control, quiet, medium, and loud) and the type of stimulus (three levels: generic song, the listener’s favorite song, and a spoken podcast). A secondary analysis was conducted to investigate the association between SNR and the output level of the personal audio system under the levels of background noise. This was conducted using Pearson correlation coefficient.

RESULTS

Questionnaire analysis

A total of 38 responses were collected from the short questionnaires administered to bus commuters. The ages of the participants ranged from 16 to 38 years (M = 21.8, standard deviation = 5.2). Of the respondents, 50% reported that they had a vague idea about NIHL as a health issue, whereas 24% said they were aware of NIHL and 24% said they were not aware of NIHL. In total, 68% did not believe their personal audio system was set to a level that could result in NIHL, while close to a third (32%) thought it could be. When asked whether they sometimes perceived the noise level of their personal audio system as being too high when they got off the bus (and the background noise level dropped), 26% said they did, 53% reported that they sometimes did, and 21% said that they did not think the sound level was too high. Most of the respondents (94%) reported listening mainly to music, 3% mainly to podcasts, and 3% listening approximately equally to both music and podcasts. The commuters who completed the survey reported spending on average 6.3 hours commuting on public transport each week and, while commuting, spending an average of 5.2 hours listening to their personal audio systems, equivalent to about 1 hour every working day of the week.

Sound pressure level setting behavior

Figure 1 shows the average and peak sound levels for each of the bus commutes, ranked by decreasing average level. Sound levels ranged from 71.3 to 78.3 dB LAeq, and peak levels ranged from 115.0 to 125.6 dB LCpeak. On this basis, the average levels used for the bus background noise were 71.7 dBA for “soft” and 77.1 dBA for “medium.” In addition, a level of 81.2 dBA was assumed for “loud.” This level was not observed in Auckland, but is consistent with those of noisy city buses found elsewhere, as noted in the international literature for large cities in developing countries.[56]

Figure 1

Sound pressure levels (dB Leq and dBCpeak) for commutes over a period of 20 minutes across a range of different bus routes, ranked by decreasing average sound level (original).

Table 1 summarizes the preferred personal audio system sound levels identified by the individual participants for the four levels of background noise. The bold values in Table 1 are all of the combinations of participants and sources (podcast, generic, or favorite song) that would have resulted in an exceedance of the daily noise dose if in an occupational setting, assuming the participant had a commute of 1 hour a day and no significant exposure to occupational or other environmental noise throughout their day.

Table 1

Sound levels for each of the participants subjected to each combination of levels of background noise and types of music/podcast

Subject	Control(dBA)			Quiet(dBA)			Medium(dBA)			Loud(dBA)
	Generic song	Favorite song	Podcast	Favorite song	Generic song	Podcast	Generic song	Favorite song	Podcast	Generic song	Favorite song	Podcast
1	64	68	67	83	89	90	85	91	93	90	92	99
2	65	66	65	73	71	81	74	82	81	82	83	83
3	79	83	80	86	93	82	91	94	87	93	93	94
4	66	74	70	79	78	81	80	78	84	84	84	91
5	70	83	60	84	93	78	95	93	80	90	93	81
6	75	75	77	86	85	90	91	88	90	89	91	91
7	65	62	65	86	82	82	84	83	85	88	84	89
8	62	48	63	83	80	84	81	87	93	89	88	96
9	84	84	73	88	94	86	86	99	93	100	96	94
10	77	85	74	86	95	92	88	93	99	89	199	101
11	82	78	82	86	86	87	91	91	95	89	96	90
12	88	81	82	90	85	89	90	94	90	96	92	90
13	110	108	102	111	110	112	111	111	108	111	111	109
14	87	88	79	92	95	92	91	94	92	92	96	93
15	75	68	73	89	81	91	92	90	95	92	92	94
16	82	80	68	86	92	86	96	93	95	100	93	96

Bold font indicates an exceedance of the daily noise dose (original).

Figure 2 shows the average sound output levels for each background noise intensity condition and each sound type. In general, the preferred personal audio system output sound levels increased systematically with increasing level of background noise (F(3,45) = 73.849, P < 0.001) [Figure 2]. There was also an interaction between the bus noise level and the type of entertainment, in that the listening level for the podcasts was lower in the control condition but increased more rapidly than the levels of the music conditions with increasing level of background noise (F(6,90) = 3.315, P = 0.005) [Figure 2].

Figure 2

Average sound pressure level when subjected to no background noise (“control”), and noise from a “quiet,” “medium,” and “noisy” buses when listening to a different types of music/podcast. Error bars are the standard error of the mean.

The data suggest that when traveling on a quiet, medium, and loud bus and listening to generic music, 6%, 19%, and 25% of commuters exceed their daily allowable noise dose. When listening to their favorite song, these figures are 25%, 31%, and 38% and when listening to a podcast, the figures are 6%, 31%, and 50%. Note that participant 13, an outlier among the participants, had a preferred listening level that was comparatively very high (always in excess of 102 dB), irrespective of the level of background noise.

Speech-in-noise ratio

Figure 3 shows the SNR score for each of the participants as a function of the average sound level selected for each listening condition. This figure suggests that there is no significant relationship between the preferred volume settings in either the presence or absence of background noise (at any of the levels) and the SNR score (r = 0.30, P = 0.9 and r = 0.40, P = 0.142 for the control case and in the case of medium level of background noise, respectively, when listening to a podcast with the outlier excluded); the majority of the participants had normal SNR scores based on the scoring criteria provided by Etymotic Research Inc (i.e., SNR score between 0 and 3 dB). However, one participant, participant 13, was found to have a moderate SNR loss despite having normal pure-tone audiometric thresholds in quiet. Associations between SNR and listening levels did not become significant after removal of this case as an outlier.

Figure 3

Sound output level as a function of the speech-in-noise ratio (SNR) score (in dB) of study participants for all four conditions of background noise (including “control”).

DISCUSSION

The aims of this study were to measure the noise levels on Auckland public buses and determine whether the level of background noise they produce influenced personal audio system users’ volume-setting behavior when listening to either music or podcasts. A written questionnaire administered to bus commuters gave an indication of the approximate daily travel time and the attitude and knowledge regarding NIHL among personal audio system bus commuters.

The average sound pressure levels recorded on the buses ranged from 69.2 to 77.8 dBA. This suggests that the noise level experienced is unlikely to cause large audiometric losses, even over 8-hour shifts and many years of working in the role for the drivers of the buses, nor for any passenger traveling in them for their daily commute on the basis of the engine noise alone.

On the other hand, personal audio system use would add a potentially significant amount of extra noise-exposure burden to occupational noise experienced, and in many cases, it would cause commuters to exceed the legal workplace noise exposure without even being exposed to sound in their jobs. The proportion of bus commuters with personal audio systems and listening to music that would exceed 85 dB LAeq, 8 hours during 1 hour of daily commuting was found to increase with increasing level of background noise; with levels reaching 12% to 38% commuters on a “medium” bus. Even on a “quiet” bus, about 6% to 25% of commuters would have personal audio systems exceeding 85 dB LAeq, 8 hours during commuting alone and depending on what they were listening to. For a “loud” bus, there levels increased to 25% to 44% depending on what is being listened to. One study[15] investigated volume settings in the presence of different levels of background noise and found that for background noise levels of 70 to 80 dBA, sound levels were set to 88 dBA on average. Even for the “medium” bus in our study, the average sound levels were found to be about 90 dBA. In this study,[15] real-ear probes were used to record the sound levels whereas in the current study, an artificial ear was used. However, in contrast with the present study, only a small proportion of the participants owned their own personal audio system or used one regularly, perhaps a reflection of the time at which the study was carried out, before the use of mobile phones for listening became highly prevalent. A related study[16] presented the results investigating background noise consisting of white noise, speech babble, as well as bus or train noise. For a noise level of 75 dB SPL, it was found that 30% of bus commuters would exceed the workplace daily allowable noise dose when listening to their favorite song. This is similar to the results found in our study where on average 31% of individuals would have exceeded their daily allowable noise dose, whereas commuting on a bus of medium noise level. Again, this study[10] used real-ear probes to record the SPL experienced by each of their participants. In this case, though, all of the participants owned their own personal audio systems.

Given that people who work in moderately noisy occupations (e.g., 70–80 dB LAeq, 8 hours) are not protected by workplace safety legislation, even small amounts of noise exposure during commuting could be significant and the high levels revealed in this research imply that the noise exposure standards, which were developed in a society which did not have personal audio systems, may need to be revised.

In the present study, one participant chose a very high volume setting, irrespective of the background noise level or the entertainment types chosen (podcast or music). At the levels selected by this one individual, even 1 minute of exposure is greater than 85 dB LAeq, 8 hours. Among all of the participants, this one participant was the only one to also have a moderate SNR loss. With the small sample size, no conclusions can be made but it suggests in might be possible that there is a relationship between SNR score and preferred listening levels, though the direction of the effect is impossible to know in a cross-sectional study. This is an area of study warranting further investigation.

The buses used in this study were all diesel buses as these make up the majority of Auckland’s bus fleet. In the future, it would be helpful to look at hybrid and electric buses as these can be expected to be quieter. Also, advances in headphone technology have been made since the data for this study were collected. The more widespread use of noise-canceling technology and earphones that occlude the ear canal to block out background noise can be expected to help considerably in limiting the impact of background noise on preferred volume setting.

At the same time as sound level measurements were made, a short questionnaire was administered to young adult bus commuters who were using personal audio systems. The questionnaire revealed that most people listened to music rather than podcasts while commuting. The survey also revealed that, on average, commuters spend about an hour a day on the bus. This estimate was used to predict the average daily dose in relation to the preferred volume settings selected. In reality, some people will commute for significantly longer periods and some for shorter periods.

A limitation of the current study was the use of an artificial ear when measuring the sound level output from the personal audio system. This was carried out to help with repeatability but would not have accounted for the person-to-person variability in sound levels that would have been observed with real-ear measurements. In addition, the small number of participants involved in the lab testing was a significant limitation of the study. A larger sample size would have been more representative of the listening behaviors and the risk posed to hearing within the population. Nonetheless, the results give an indication of the potential extent of the issue.

CONCLUSION

The noise levels experienced by commuters on city buses in Auckland are below those required to reach unsafe levels for typical commutes of 1 hour per day (71.3–78.3 dBA). However, for many who use personal audio systems while commuting by bus, the volume settings chosen are such that the safe noise dose is exceeded regularly. Further education is needed in relation to NIHL as results from the questionnaire suggest that the majority of young people are unaware that there is a risk associated with the use of personal audio devices with respect to hearing. Moreover, current noise regulations in New Zealand are focused on occupational noise, highlighting the need for solutions to be developed for addressing recreational noise exposure, as well to help protect the hearing of the population at large.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.