Preferred listening levels of mobile phone programs when considering subway interior noise.

Today, people listen to music loud using personal listening devices. Although a majority of studies have reported that the high volume played on these listening devices produces a latent risk of hearing problems, there is a lack of studies on "double noise exposures" such as environmental noise plus recreational noise. The present study measures the preferred listening levels of a mobile phone program with subway interior noise for 74 normal-hearing participants in five age groups (ranging from 20s to 60s). The speakers presented the subway interior noise at 73.45 dB, while each subject listened to three application programs [Digital Multimedia Broadcasting (DMB), music, game] for 30 min using a tablet personal computer with an earphone. The participants' earphone volume levels were analyzed using a sound level meter and a 2cc coupler. Overall, the results showed that those in their 20s listened to the three programs significantly louder with DMB set at significantly higher volume levels than for the other programs. Higher volume levels were needed for middle frequency compared to the lower and higher frequencies. We concluded that any potential risk of noise-induced hearing loss for mobile phone users should be communicated when users listen regularly, although the volume level was not high enough that the users felt uncomfortable. When considering individual listening habits on mobile phones, further study to predict total accumulated environmental noise is still needed.

Introduction

It is acknowledged that exposure to loud noise over a long duration can cause noise-induced hearing loss (NIHL).[1] Although workers have suffered from NIHL due to workplace noise in the past,[2] more recently, recreational noise from entertainment sources has been proved to be one of the factors producing NIHL.[3] It has also been widely reported that children and youth who did not have the experience of occupational exposure to noise show NIHL because they may be constantly exposed to noise from home machines and applications, habitual loud music, and high traffic noise.[45] Of all those noise activities, the level of noise exposure from music on a portable listening device shows a high positive correlation to the density of population and degree of urbanization.[6] Indeed, organizations in some countries have recommended criteria for permeable noise exposure level to prevent NIHL in the past. For example, European countries have fixed the maximum volume level of portable listening devices to 100 dBA beginning in 2002.[7] The U.S. Occupational Safety and Health Administration has also regulated laws, indicating that NIHL can occur if people listen to music for 2 h or longer at 100 dBA.[8] Some countries have even recommended the use of portable listening devices for less than 1 h per day at no more than 60% of maximum volume because the higher levels can cause damage to the human auditory system.[9]

As technology has developed, the mobile phone has taken on fancier and more complex functions, including moving picture, games, and sound applications. In South Korea, where the information technology (IT) industry is a world superpower, the marketing of mobile phones has significantly increased year by year and new devices continue to be released with a variety of attractive features. As a result, people there use a wide range of mobile device applications taking full advantage of digital imaging and sound effects. Currently, a majority of Korean people listen to music and watch movies outdoors using the application programs on their mobile phones. The trend is for older as well as younger people to enjoy their phones. A recent study by Kim (2009) found that 60% of 1,480 respondents use their mobile phone programs for 1 h or longer per day and 10.8% of the respondents use them for more than 3 h every day to listen to audio.[10] The Korean government and phone manufacturers voluntarily agreed in 2014 that the maximum volume level of the mobile phone should not exceed 100 dBA when sound files are played.[11]

Nevertheless, despite such positive regulatory interest in sound output levels, a large number of mobile phone users have responded by saying that they usually increase their volume levels in noisy situations, thus ignoring the recommended output limit for sound levels on their phones.[12] Further, most of the research conducted to date has focused on young people listening to music and has not considered the characteristics of users from different age groups who have benefited from the fast penetration of the mobile phone market and the multiple sound features of new application programs available on their devices. Because the previous studies have emphasized only a uniform standard of noise exposure at the level of 100 dBA to the general public, they have failed to provide realistic and practical criteria for users overall. In other words, while many studies have reported that a high volume level on listening devices carries a latent risk of hearing loss, researchers have neither accurately analyzed that volume level nor provided scientific evidence-based precautions to prevent NIHL in users.[13] Indeed, there is a lack of studies on “double noise exposures” such as environmental noise plus recreational noise (i.e., practical volume levels for noisy situations).

As subway noise is of the more common environmental noises, Gershon et al. (2006) measured subway noise levels in New York City, showing that the noise level on subway platforms was about 85.7 dBA, with subway interior noise ranging 84~112 dBA.[14] The authors concluded that subway passengers are at a slight risk for NIHL from exposure to subway noise alone even without any exposure to secondary recreational noise, such as their use of personal listening devices during and after commute. As subway interior noise is caused by concrete-based rail tracks and a tunnel structure,[15] its interior sound pressure level as heard by passengers can be as high as 60~105 dBA in the low frequencies and will vary based on the speed of the subway train.[16] As a result, passengers may have to increase the volume levels on their listening devices to overcome the background noise from the subway interior.[12] When including preferred listening levels as part of an ambient noise environment,[17] it would appear that age, gender, and other sociological characteristics may also contribute to the levels of sound exposure and risk with the use of these devices.[4]

The purpose of the present study was to measure the preferred listening levels of mobile phone programs considering subway interior noise for commuters, while also considering various age groups, mobile application programs, and affecting frequencies. The study hypothesized that younger listeners would prefer high volume levels and that programs with a playback sound function would characterize a listener's habits in terms of using higher or lower volume levels. We also expected that frequency specificity would be indicated.

Methods

Participants

Seventy-four subjects (37 males and 37 females), who were randomly recruited through posted flyers at community centers, participated in this study, where they were placed in five age groups (ranging from 20s to 60s). The average age of each group was as follows: 23.5 years (std: 1.83) for the 20s group (8 males and 8 females); 33.36 years (std: 3.08) for the 30s group (8 males and 6 females); 45.08 years (std: 3.07) for the 40s group (5 males and 8 females); 53.88 years (std: 2.5) for the 50s group (8 males and 8 females); and 64.80 years (std: 2.43) for the 60s group (8 males and 7 females). All participants reported a negative history of head or neck abnormalities, ear surgery, otologic disease, and/or head trauma. They also passed the normal criteria for hearing screen-ing to ensure an A-type of tympanogram and had normal hearing sensitivity as a function of their age in each ear at 250~8,000 Hz[18] and air-bone gaps no greater than 5 dB hearing level (HL). All participants also signed an informed consent form before the experiment and received $30 for completing the experiment. The experimental procedure was reviewed and approved by the Institutional Review Board of Hallym University (HIRB-2013-039).

Stimuli

To generate subway interior noise, nine different subway lines inside Seoul in Korea (Lines 1~9) were measured and recorded using a sound level meter (Type #2250, Brüel and Kjær, Nærum, Denmark) twice for 2-h periods in the morning and evening during commuting time. The average for this measured subway interior noise was 73.45 dB LAeq (std: 2.57). To learn preferred listening levels for an application program for mobile devices, three of the most popular programs were chosen: Digital multimedia broadcasting (DMB), music, and an online game. Each program had two different kinds of sound characteristics. For example, the DMB program consisted of one typical drama and one entertainment show. One music program was a medley of ballads, while the other was a medley of dance music. To select two online games, we ranked them in order of popularity and picked two of these, while also considering their heterogeneous background sounds [Figure 1].

Figure 1

The laboratory scene for the simulation with 5.1 channel speakers generating subway interior noise (top), a tablet PC, earphones, a two-in-one plug (left bottom), a 2cc coupler, artificial ear, and a sound level meter to measure each subject's preferred volume (right bottom)

Experimental procedures

After completing the hearing screening tests, each subject listened to three application programs in random order for 30 min each using a tablet personal computer (PC) (iPad, Apple Co., Cupertino, CA, USA) and an earphone (model # MDR-E9LP, Sony Co., Tokyo, Japan), while 5.1 channel speakers presented the subway interior noise at 73.45 dB in the same room [Figure 1, top]. One side of the earphone called an ear bud was connected to the tablet PC, and the other side plug was connected to an analysis system via a two-in-one plug [See left bottom in the Figure 1]. The analysis system consisted of a 2cc click-on coupler (Type #4946, Brüel and Kjær, Nærum, Denmark), an artificial ear (Type #4153, Brüel and Kjær, Nærum, Denmark), a sound level meter, and a PC. In sum, each participant's earphone volume level was measured by a sound level meter via the 2cc coupler and the artificial ear [Figure 2, right bottom]. At the same time, the researcher analyzed the preferred volume for the earphone using computer software (BZ-5503 Measurement partner suite, Brüel and Kjær, Nærum, Denmark), while the volume level was also checked by an inter- and intra-program.

Figure 2

Mean volume levels for LAFmin, LAeq, and LAFmax analyses as a function of three mobile application programs and five different age groups

Statistical analysis

A repeated measures analysis of variance (ANOVA) was used to confirm the significant mean difference of LAeq (equivalent level of A-weighted decibels), LAFmin (maximum level of fast time and the A-weighted decibels), and LAFmax (minimum level of fast time and A-weighted decibels) measures. At each measure, the subject variable was age (five groups) and gender (male and female), and within each subject, the variable was the mobile application program (three programs) and the number of analyzed frequencies (33 frequencies). When necessary, a Bonferroni correction was used as a multiple comparison post-hoc test. The criterion used for statistical significance was p < 0.05 in this study. Statistical analysis was performed using SPSS software (ver.20, IMB Co., Armonk, NY, USA).

Results

LAeq analysis

ANOVA confirmed a significant main effect of age [F (4, 64) = 3.328, p = .015], demonstrating that the 20s group had a higher volume level (mean = 51.920 dB, std = 1.171) than the 50s group (mean = 46.752 dB, std = 1.171) [See dark-shaded bar graphs in Figure 2]. There was no significant difference of means between males (mean = 50.281 dB, std = .784) and females (mean = 48.678 dB, std = .775) [F (1, 64) = 2.117, p = .151]. In the mobile application, three programs showed a significant mean difference [F (2, 128) = 162.669, p < .0001]. A Bonferroni post-hoc test showed there was no significant difference between music (mean = 52.816, std = .660) and DMB (mean = 52.387, std = .547) (p = 1.000), but the game volume (mean = 43.235, std = .732) was significantly lower than the music volume and its DMB (p < .0001) [Dark-shaded bar graphs in Figure 2]. Another main effect was that the frequency showed a significant difference of mean [F (32, 2048) = 12016.264, p < .0001] while having an interaction effect based on age [F (128, 2048) = 1.810, p < .0001] and specific program [F (64, 4096) = 91.925, p < .0001]. A one-third octave, among a total of 33 frequencies, 315Hz (mean = 77.809, std = .607) was the highest intensity, and 12.5Hz (mean = -3.122, std = .518) was the lowest intensity. In particular, the middle frequency regions (i.e., 250~2,000 Hz) showed meaningful age and program effects, which meant that the 20s groups used the highest volume level and their DMB was at a slightly higher level than the music [See the dark solid lines in Figure 3] in the middle frequency, compared to the very low and/or very high frequencies, respectively [Figures 2 and 3].

Figure 3

Mean volume levels for LAFmin, LAeq, and LAFmax analyses and three mobile application programs as a function of frequency

LAFmax analysis

In the LAFmax analysis, ANOVA also confirmed a significant main effect of age [F (4, 64) = 3.500, p = .012], program [F (2, 128) = 172.473, p < .0001], and frequency [F (32, 2048) = 5912.966, p < .0001], except there was no gender difference [F (1, 64) = 1.838, p = .180]. As a result of a Bonferroni multiple comparison, the 20s group showed a significantly higher volume level (mean = 70.092 dB, std = 1.169) than did the 60s group (mean = 64.248 dB, std = 1.21) [Slash bar graphs of Figure 2]. DMB (mean = 71.629 dB, std = .599) was significantly higher than music (mean = 68.202 dB, std = .647) (p < .0001) and game (mean = 60.011 dB, std = .738) was significantly the lowest (p < .0001) [Slash bar graphs on Figure 2]. In terms of frequency effect, 400 Hz (mean = 93.474 dB, std = .641) was the highest intensity, and 12.5 Hz (mean = 18.039 dB, std = .570) was the lowest intensity [Dashed lines on Figure 3]. There was no interaction effect.

LAFmin analysis

In the LAFmin analysis, there was no significant difference in mean in the five age groups [F (4, 64) = .892, p = .474]. That is, there was about a 1-dB difference between the 20s group that had the highest volume level at 17.559 dB and the 60s group that had the lowest volume level at 16.453 dB. As in the LAFmax analysis, there was no significant difference in the gender effect [F (1, 64) = .931, p = .338]. However, three mobile application programs did show a significant mean difference [F (2, 128) = 68.919, p < .0001], while DMB (mean = 19.476 dB, std = .390) was significantly higher than music (mean = 16.967 dB, std = .324) (p < .0001), and game (mean = 14.599 dB, std = .297) was significantly the lowest (p < .0001). Frequency effect showed a significant difference [F (32, 2048) = 4052.250, p < .0001], and 315 Hz (mean = 47.988 dB, std = .542) was the highest intensity and 12.5 Hz (mean = -30.941 dB, std = .505) was the lowest intensity. An interaction effect between frequency and the program was also found [F (64, 4096) = 35.517, p < .0001]. DMB showed the highest volume level across all the analyzed frequencies, and music showed a higher level than the game in the middle frequencies, from 63 Hz to 3,000 Hz, while showing a level similar to the game in the very low- and/or very high-frequency regions [Dated lines on Figure 3].

Discussion

The purpose of the present study was to measure the preferred volume level for mobile application programs when used along with subway interior noise. The LAeq analysis showed that 1) the 20s group listened to three application programs at significantly 5 dB louder than the 50s group did; 2) DMB and music programs were significantly 9 dB louder than the game, and yet they were not a significant difference; 3) there was a significant difference for the measured 33 frequencies with 315 Hz the highest sound pressure level as 77.81 dBA, whereas 12.5 Hz was the lowest level.

In the LAFmax analysis, 1) the 20s group listened to three application programs significantly 5 dB louder than did the 60s group; 2) the volume level of DMB was significantly 5 dB higher than that for music, which was 8 dB higher than that for the game; 3) 400 Hz and 12.5 Hz were the highest and lowest sound pressure levels, respectively. On the other hand, the results of the LAFmin analysis showed that 1) there was no significant age effect; 2) DMB was at a significantly higher volume level than the music, and music also was at higher volume level than the game. However, the levels of the three programs were not much different statistically; 3) 315 Hz and 12.5 Hz were the highest and lowest sound pressure levels, respectively. The three analyses did not show any interaction and/or gender effects.

In terms of an age effect, the younger age groups were used to listening to louder music, although their hearing sensitivity was generally better than that of the older groups. It is estimated that over 90% of college students own a personal listening device,[19] and 5~10% of these device users regularly listen to music at high levels for periods of time that can put them at risk for NIHL.[4] One of the reasons why this group enjoys listening to loud music is that for them, wearing earphones represents a modern, urban lifestyle and leisure for young people. A second reason is that music or sound stimulated by the earphones of personal listening device changes their emotional state to either one of relaxation and/or excitement, depending on the rhythm.[4] To prevent the development of NIHL, many researchers have suggested that specific public education on appropriate volume levels and exposure time is necessary for the younger generation. In contrast to our expectations, there was no difference in the preferred volume levels between male and female users. The preferred volume level is likely explained by personal characteristics rather than by gender. This finding is further supported by previous results[412]

Of the three application programs, DMB showed the highest preferred volume level. Because partial subtitles were shown as keywords in the entertainment program, we expected that the users did not need a high volume. However, conversations between various characters in a dramatic or an entertainment program needed more concentration, resulting in an increase in the volume level. In a study by Vogel et al. (2008), those participants explained their motives for playing sound at a high volume level. They wanted to increase the volume level to reduce background noise and to hear the sound well. Especially when a favorite song was played or when they wanted to sing along as they listened, they would increase the volume level.[13] That is, their preference and concentration on the program needed a higher volume level for the listener, which also applied to our results of the music program. However, among the kinds of music, there was no significant difference in volume level when the users listened to a medley of ballads and to dance music, according to our data. Interestingly, unlike our early expectations, the mobile game did not need a high volume level when the users played the game, regardless of the presentation of background music. There appeared to be higher dependence on visual effects than on the audio effects.

In terms of consistency of using a frequency, our results showed that the middle frequencies of 100~3,000 Hz needed more energy than did other frequency regions when users listened to programs on their mobile devices. This finding might indicate that sound specificity for a mobile phone usually covers conversation frequency for human communication.[20] Although many movements for upgrading both a quality and the intelligibility of speech in the telephone system might extend bandwidth to high frequency,[21] most systems in use today have frequency range of 300~3,400 Hz, as several researche studies have mentioned.

Overall preferred volume levels were not as high as the indicated volumes in previous studies. Hodgetts et al. (2007) showed that listeners had a mean of preferred listening level for music at 76.0 dBA in quiet, 83.7 dBA in a 70 dBA multitalker babble environment, and 85.4 dBA with 70~80 dBA street noise.[22] Such a discrepancy helps to explain that our preferred volume levels did not need to be as high, because some of the programs in our experiment included video clips with a story. Although our results were well below the noise exposure levels commonly cited as an acceptable risk for hazardous noise exposure, we did consider them as being above the level considered to carry negligible risk.[23] Further, Vogel et al. (2008) pointed out that a plausible explanation for this underestimation might lie in a more gradual development of hearing loss and because most people with mild high-frequency hearing loss are actually unaware of their growing impairment.[13]

In terms of future, needed study, having more subjects participate in a study of detailed experimental design can produce a stronger generalization. In addition, we need to develop a predicting formula for NIHL to prevent it beginning at a younger age. To promote a proper sense of perceived vulnerability, specific information on dangerous decibel levels should be provided users of mobile equipment. This information should also include suitable exposure time.[13] Specific evidence-based and theory-based studies and preferably more longitudinal studies should be conducted on the current mobile phone generation to develop the most effective interventions to prevent NIHL. We expect that such studies will be able to assess useful correlates in greater depth, including the roles of healthcare specialists for monitoring and mobile phone user perceptions of risk-taking, and those future consequences.[13]

Conclusion

Based on age and application program use by mobile phone users, this study presented and analyzed different preferred volume levels along with subway interior noise depending on frequency. Based on these study results and their analyses, we expect that a potential risk to hearing loss for mobile device users is present, and that users should be better informed about that risk. We suggest continued study into this issue to predict more precise environmental noise accumulation levels based on individual listening habits for different ages.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.