Work-related communicative profile of radio broadcasters. A case study

Abstract

Purpose:

To explore the short-term effect of work-related voice use on voice function, and noise exposure on hearing function among radio broadcasters.

Method:

A one-week follow-up study with the participation of two radio broadcasters was conducted. Participants were monitored at the beginning and at the end of the working week. Pre-monitoring assessment on Monday (baseline measure) and post-monitoring assessment on Friday (follow-up measure) were performed to identify short-term effects of work-related conditions on voice and hearing function among radio broadcasters.

Result:

Changes in fundamental frequency post-monitoring at the end of the work week may be an indication of work-related vocal fatigue. Changes in the distribution and standard deviation of SPL during the monitoring from Monday to Friday may indicate control of the vocal loudness as a strategy to reduce vocal effort during broadcasting. During a one-week follow-up, noise conditions during radio broadcasting were below occupational exposure limits and without noticeable consequences on hearing function.

Conclusion:

The work-related communicative profile of radio broadcasting, from this pilot study, suggests that although vocal demands in terms of vocal load may differ among broadcasters, the work-related conditions of broadcasting may play a role on vocal function among these occupational voice users. Concerning hearing function, our results indicate that occupational noise exposure be represented minimal risk for hearing problems but the consequences of long-term noise exposure on hearing mechanisms may yet occur. Future studies with bigger sample sizes are warranted to confirm our results.

INTRODUCTION

Previous studies have reported a higher occurrence of voice disorders among occupational voice users, partially associated with work-related conditions, such as prolonged periods of work-related voice use (vocal load) (1–3). Voice monitoring during working time has shown elevated voice use among this population. It has been found that vocal folds vibrate up to 15% of the time at work among call-center workers (4) and up to 40% of the time among teachers (5, 6). Because for each vibration of the vocal folds, the vocal fold mucosae experiences both shear stresses and collision forces, the prolonged vocal fold vibration during occupational voice use (vocal load) may lead to repeated injuries which the body repairs through connective tissue thickening (7). In addition to the vocal load linked with the occupational voice use, there are other work-related factors that have been previously associated with the occurrence of voice disorders among occupational voice users. A significant association between environmental factors (such as noise levels and acoustic conditions) and self-reported voice symptoms, such as vocal fatigue, vocal effort, hoarseness, and tired voice has also been reported (8–12). For example, previous studies have reported that teachers whose workplaces had high background noise levels had a higher likelihood of reporting voice complaints compared with their colleagues who were not occupationally exposed to these conditions (13, 14). A similar relationship between background noise levels and voice complaints has also been reported among call-center operators (15). It has been shown that speakers tend to raise their vocal levels and fundamental frequency under noisy conditions (Lombard speech) (16), which can cause higher vocal effort and vocal fatigue (12, 17–19) increasing the risk for voice problems.

While many studies have examined the influence of work-related factors on voice production among occupational voice users generally, few have looked at the effect among radio broadcasters specifically. Previous research has reported that, among occupational voice users, radio broadcasters are among those who have higher vocal demands in terms of voice quality, but moderate demands on vocal load (2). Nevertheless, there is a dearth of studies on the occurrence and work-related factors of voice symptoms that could impact voice quality, such as vocal fatigue. This information is of high interest considering the high demand for radio media. A recent survey among 9,100 participants from selected countries worldwide reported that more than 60% of respondents listened to the radio at least once a week (20). Moreover, the average daily time spent per individual listening to the radio in the United States in 2016 was 104 minutes (21). This means that, on average, people spend almost two hours per day receiving information from a person that they just know by voice. Although several reasons may influence the selection of one radio station over another, voice quality is among the most important. Therefore, radio broadcasters have a high demand in terms of voice quality for the performance of their occupational duties.

Additionally, since the voice is the primary tool for delivering information and hearing is the main tool for receiving information and for adjusting the production of voice, the breakdown of either mechanism negatively impacts the work process of occupational voice users. Nevertheless, there has been limited research exploring how noise exposure in the workplace may impact the hearing of occupational voice users in general, and specifically, among radio broadcasters.

One of the most important consequences of noise exposure is noise-induced hearing loss (NIHL). NIHL is a leading occupational disease, a major contributor to the development of age-related hearing loss, a significant cause of disability, and a major cost to society (22). Its first audiologic sign is a “notching” of the audiogram at the high frequencies (usually 4000 Hz) with recovery at 8000 Hz (23). NIHL is the result of continuous or intermittent noise exposure during a certain period of time, and usually develops slowly over several years. Since NIHL is the result of noise exposure, several studies have been developed on the relationship between occupational noise exposure and NIHL. Most of these studies have focused on industry, which have resulted in permissible exposure values (PEL) established by the Occupational Safety and Health Administration (OSHA). These PEL are defined as the legal limits for exposure of an employee to a substance, agent or condition, such as loud noise (OSHA 1910.95). For noise, OSHA´s PEL is 90 dB(A) for all workers for an 8 hour day with a 5 dB(A) exchange rate (when the noise exposure is increased by 5 dB(A), the time of exposure is cut in half) (24).

Regardless of this well-known standard, there are many non-industrial occupations exposed to less noise burden, yet without established limits, and are at risk for hearing and other communicative problems; this is the case of radio broadcasters (25). For example, previous studies among teachers, who are regular occupational voice users, have reported mean values of noise levels in occupied classrooms ranging from 72 dB(A) (14) to 76 dB(A) (26). Although according with OSHA 190.95, these values do not exceed the permissible exposure limits as currently defined, these noise levels inside the classrooms over time might affect the communication process, leading to voice disorders (11, 14). Lindblad et. al., (2014) reported that teachers showed comparable results as industrial workers for speech recognition in noise, which may suggest lesions in the inner hair cell area due to exposure to sudden loud sounds, such as screaming of the children. Although no previous studies on this topic were found among radio broadcasters, these results among teachers indicate that occupational exposure to unfavorable conditions for communication may be linked with a higher risk for developing voice disorders and hearing problems.

Therefore, this study was designed to answer the following research questions: 1) Is there a short-term effect of work-related voice use on voice function among radio broadcasters? and 2) Is there an effect of work-related noise exposure on hearing function among radio broadcasters? To answer, we conducted a one-week follow-up study with the participation of two broadcasters. Although the sample size is small, the results of this study can give us initial insights on the work-related communicative profile of this occupational group.

METHODS

Design and Participants

This one-week follow-up study was performed in May of 2017. Follow-up studies (also called longitudinal, concurrent, incidence, cohort, panel, prospective study) are designed to observe an individual, group, or a population “at risk” (exposed to a factor hypothesized to influence the occurrence of a given outcome) over a period of time to assess changes in health status or health-related variables (outcome) related with the “associated factor” (exposure) (27). In this specific study, broadcasters (population of interest) were followed over one week (follow-up period) to assess changes on voice and hearing function (outcome) related with occupational voice use and occupational noise exposure (associated factor). Participants were monitored at the beginning and at the end of their working week. Since none of the participating broadcasters perform their radio programs during the weekends, these two point-measurements will explore, for the very first time, changes on voice function and hearing function associated with the occupational voice use and noise exposure during one week of broadcasting. After approval from the Michigan State University Human Research Protection Program, two radio broadcasters (one male and one female) from a public radio station participated in the study. Both participants were native speakers of American English, with no voice or hearing problems during the week of the monitoring. The female broadcaster (46 years old) is part of a morning news program, which lasts for 4 hours each weekday morning, while the male broadcaster (26 years old) presents the sports news at noon each weekday for one hour. During the morning news, the female broadcaster shares the show with other co-hosts; therefore, she is not speaking all the time. During the sport news, the male broadcaster is the main host.

Data collection procedures

Pre-monitoring and post-monitoring assessment

After giving written informed consent to participate in this study, participants went under a pre-monitoring assessment. This pre-monitoring assessment was performed with the aim of characterizing the voice quality and hearing function before the occupational exposure associated with radio broadcasting. The pre-monitoring assessment included three elements. Firstly, broadcasters filled out a questionnaire, consisting of four sections: (1) 9 questions on socio-demographics (e.g. age, gender and education), native language, and history of hearing or speech disorders; (2) 16 questions on working conditions (e.g. days per week of broadcasting, hours a day of broadcasting, physical conditions of the workplace); (3) 13 questions on the occurrence, severity and frequency of voice symptoms; and (4) 17 questions on the occurrence, severity and frequency of hearing problems. Secondly, they went under a hearing screening, which was performed with the Audiometer MADSEN Orbiter 922 Version 2 (Otometrics A/S, Taastrup, Denmark) in a double wall sound isolation booth (2.5 × 2.75 m and h = 2 m) with a mid-frequency reverberation time (RT20) equal to 0.05 seconds and a background noise equal to 20 dB(A). Thirdly, they recorded a speech sample. The speech sample consisted of reading the first six sentences of “The Rainbow Passage” (28), equal to about 30 seconds of speaking in a conversational pitch and loudness. The speech samples were recorded in a double wall sound isolation booth (2.5 × 2.75 m and h = 2 m) with a mid-frequency reverberation time (RT20) equal to 0.05 seconds and a background noise equal to 20 dB(A) using a M80 omnidirectional head-mounted microphone (Glottal Enterprise, Syracuse, NY, USA) placed at 5 cm from their mouth, which was connected to a TASCAM DR-40 Linear PCM Recorder (TEAC Corporation, Tokyo, Japan). The digital recordings (44,100 Hz) were transferred to a personal computer (PC) running Praat Version 6.0.31 (29). Settings for calculation of the fundamental frequency (Fo) in Praat were: pitch range between 70 Hz and 450 Hz for the female broadcaster, and between 50 Hz and 350 Hz for the male broadcaster. In addition, participants produced a sustained vowel /a/ in a conversational pitch and loudness. The vowel was used for calculation of the Harmonics-to-Noise ratio (HNR). HNR was included because it is a measure that reflects the periodicity of vocal fold vibration, and therefore, the harmonicity of the voice. Previous research had shown that HNR is a sensitive index of vocal function (30) and a good measure of voice quality (31, 32).

After the radio shows, a post-monitoring assessment was performed. The objective of the post-monitoring assessment was to determine changes on voice quality and hearing function associated with the occupational exposure to noise and voice use during broadcasting. The post-monitoring assessment included three elements. First, the subjects were asked to complete a reduced survey consisting of two sections: (1) 4 questions on the presence, type and severity of voice symptoms after the show; and (2) 8 questions on the presence, type and severity of hearing problems after the show. Second, a hearing screening was administered. Third, a final recording of a speech sample (Rainbow passage) and the sustained vowel /a/ was made.

Since the broadcasters do not perform their radio shows during the weekends, the pre-monitoring assessment on Monday was considered the baseline measurement at the lowest short-term occupational exposure (post weekend recovery) (33). Therefore, the voice quality parameters and hearing thresholds identified during this assessment were compared with post-monitoring measurements. The purpose of this comparison was twofold: 1) to characterize changes on voice quality associated with the occupational voice use during radio broadcasting, and 2) to identify variations on the hearing thresholds associated with the occupational noise exposure during radio broadcasting.

Occupational voice monitoring

After the pre-monitoring assessment, the broadcasters were outfitted with an Ambulatory Phonation Monitor (APM, model 3200 by Kaypentax; Montvale, NJ) before starting their radio shows. The radio broadcasters were monitored during the whole duration of the shows and were instructed to behave during their radio shows as normally as possible. The APM, which has been designed in such a way that allows the participants to walk and move freely during their occupational monitoring, provides an estimation of the individual’s vocal sound pressure levels (SPLs) at a fixed distance of 15 cm from the speaker´s mouth after a calibration with a reference microphone. Device calibration, data handling, and resulting metrics of the calibration process were built into the APM software; APM manufacturer instructions for the calibration were followed. The occupational voice monitoring was carried out using an accelerometer, which measures the vibrations that occur during phonation, mounted on a silicone pad and placed at the jugular notch using surgical adhesive. The vocal parameters were estimated using an acquisition device that processed the signal into frames of 50 ms. The phonation time percentage (D_t%) is obtained through a procedure that allows the separation of the voiced and unvoiced frames. The voicing frames were determined using two different criteria: (1) a lower bound of 30 Hz and an upper bound of 400 Hz for the fundamental frequency, (2) and a voicing threshold (equal to 0.45 relative to the global maximum amplitude) and silence threshold (equal to 0.03 relative to the global maximum amplitude). A frame was rated as unvoiced if it had an intensity below the voicing threshold or a local peak below the silence threshold. The fundamental frequency (Fo) is extracted from the voiced frames with a procedure that is based on an autocorrelation algorithm. The autocorrelation analysis computes the correlation between a signal and a delayed copy of the same signal at delays corresponding to the minimum and maximum expected fundamental periods (34). This method is more accurate, noise-resistant, and robust than other methods based on the cepstrum or combs (35).

Occupational noise monitoring

Noise exposure among broadcasters may have two sources: (1) environmental noise inside the radio studio, and (2) noise from the headphones during broadcasting. Occupational noise exposure during the current study (assessed according to OSHA Standard 1910.95) was determined by two techniques. The environmental noise inside the radio studio was determined by means of an omnidirectional microphone (M2211, NTi Audio, Tigards, OR, USA) placed at around 70 cm from the position of the broadcaster during broadcasting. The noise exposure produced by means of the headphones were assessed using the Manikin or Head and Torso Simulator (HATS) technique according with the ISO 11904–2 (36). The HATS model used was the KEMAR HATS 45BC-2 by G.R.A.S.S. Sound & Vibration (Holte, Denmark). The HATS transfer function makes it possible to transform the noise level measured within the ear of the manikin into an equivalent level. To measure the noise delivered by the headset while the broadcaster was working, the input line for the broadcaster’s headset was split into two equivalent input lines. Two headsets of the same model were connected to the input lines, the first one worn by the broadcaster and the second one by the HATS. This ensured that the broadcaster and the HATS were exposed to the same levels even when the headset volume was modified (figure 1). The noise signals acquired by the microphone M2111 (NTI Audio) and the ears of the HATS (G.R.A.S.S.), were analyzed with the XL2 Analyzer (NTI Audio) to obtain the A-weighted equivalent level.

Figure 1.

Measurement setup of occupational noise exposure during broadcasting

Statistical Analysis

All analyses were performed by means of SPSS 21 (IBM Corporation, New York, USA). First, we used the Kolmogorov-Smirnov Test to assess the distribution of fundamental frequency and vocal sound pressure levels before, during, and after the monitoring. None of the variables were normally distributed (p-value=0.00). Second, bean plots were drawn to compare the distribution of fundamental frequency before, during, and after the monitoring for both broadcasters. The bean plot was chosen because it is an uncomplicated way to compare distributions by showing the frequency and density of the occurrences/observations. Each small line/dot in the bean plot represents one observation. The density shape used is a polygon given by a normal density trace and its mirrored version (37). Third, histograms were used to compare distribution of vocal sound pressure levels during the monitoring. For the vocal sound pressure levels, histograms were better than the bean plots due the density proportion of the data.

RESULTS

Occupational voice use during broadcasting

Three parameters were defined to measure the occupational voice use during broadcasting. First, phonation time percentage (Dt%) was used to represent the time during which the vocal folds vibrated (voicing time) over the total time of occupational monitoring. Second, fundamental frequency (Fo) was used to represent the average number of collision per second between the vocal folds. Third, vocal sound pressure level (SPL) was used to represent the average pressure level emitted by the participant at 15 cm from the mouth. Occupational monitoring lasted during the total duration of the radio shows. The female broadcaster´s news program was 4 hours (Monday and Friday), and the male broadcaster´s sports news program was 1 hour (Monday and Friday).

Figure 2 shows the occupational Dt% for the participating broadcasters during their radio shows on Monday and Friday. As shown in the figure, the male radio broadcaster has a higher vocal load (around 38% on Monday and 42% on Friday) compared with his female colleague (around 14% on Monday and 12% on Friday). The results indicate that during the one-hour show, the male radio broadcaster has a Dt% equivalent to the production of a “monologue”, speaking for almost the entire one-hour long broadcast; previous studies have reported Dt% from males reading a text to be between 39–54% (38). In contrast, the female radio broadcaster has a lower vocal load during the four-hour show.

Figure 2.

Phonation time percentage of participating broadcasters during their radio shows on Monday and Friday

Figure 3 shows the female broadcaster fundamental frequency (Fo) during the occupational monitoring plotted in bean plots. The female broadcaster had a slightly lower fundamental frequency mean during the monitoring on Friday (Fo= 160 Hz, SD= 38 Hz) compared with the monitoring on Monday (Fo= 171 Hz, SD= 56 Hz). The bean plots show a higher density of low Fo values (below 100 Hz) and high Fo values (above 230 Hz) during the monitoring of Monday compared with the monitoring on Friday (which is also reflected in the Fo SD). Therefore, it seems likely that the female broadcaster was producing more intonation changes during her monitoring on Monday compared with Friday.

Figure 3.

Fundamental frequency of the female broadcaster during radio shows on Monday and Friday

Figure 4 shows female broadcaster vocal sound pressure levels (SPL) during the occupational monitoring plotted in histograms. As shown in the figure, although difference in mean SPL from Monday (SPL= 82 dB, SD= 14 dB) to Friday (SPL= 83 dB, SD= 18 dB) was just 1 dB, distributions are significantly different. On Monday, the female broadcaster had a higher peak around 80 dB, whereas, on Friday, the distribution shows more dispersion without visible peaks.

Figure 4.

Vocal sound pressure levels of the female broadcaster during radio shows on Monday and Friday

Figure 5 shows the bean plots for Fo during the occupational monitoring of the male broadcaster. As shown on the figure, Fo on Friday is slightly lower (Fo= 152 Hz, SD= 45 Hz) compared with Monday (Fo= 161 Hz, SD= 47 Hz). However, in this case the density distributions are similar. Therefore, it seems likely that the male broadcaster used his voice in a similar way during both monitoring sessions.

Figure 5.

Fundamental frequency of the male broadcaster during radio shows on Monday and Friday

Figure 6 shows the histograms of the male broadcaster vocal sound pressure levels (SPL) during the occupational monitoring. The male broadcaster had lower SPL on Friday (SPL= 79 dB, SD= 13 dB) compared with Monday (SPL= 86 dB, SD= 12 dB). The figure shows that the highest peak of the distribution is slightly skewed to the right on Monday compared with Friday. In addition, there is bigger asymmetry to the right in the SPL values recorded during the monitoring on Friday compared with Monday.

Figure 6.

Vocal sound pressure levels of the male broadcaster during radio shows on Monday and Friday

Short-term effect of radio broadcasting on voice function

To assess short-term effects of broadcasting on voice function among radio broadcasters, we compared the information collected during the pre-monitoring assessment on Monday (baseline) with the post-monitoring assessment on Friday (follow-up). Participating broadcasters did not report any current voice symptoms (hoarseness, weak voice, voice loss, breathiness, strained voice, tired voice, pain in throat, itchy sensation, dry throat) either at the baseline or at the follow-up.

Concerning the short-term effect of occupational voice use on voice function, the results show a tendency to increase the fundamental frequency (Fo). As shown in Figure 7, the male broadcaster had an increase of about 16 Hz (115 Hz on pre-monitoring Monday, and 131 Hz on post-monitoring Friday), whereas the female broadcaster had an increase of about 7 Hz (140 Hz on pre-monitoring Monday, and 147 Hz on post-monitoring Friday).

Figure 7.

Fundamental frequency of female and male broadcasters during the baseline (pre-monitoring Monday) and follow-up (post-monitoring Friday) measures. The voice production corresponds to about 30 seconds of reading.

Further, Figure 8 shows that there is a different tendency between the male and the female broadcasters on the Harmonics-to-Noise ratio (HNR). The male broadcaster has a decrease in the Harmonics-to-Noise ratio (HNR) in the sustained vowel /a/ recorded at the end of the week (20 dB) compared to the beginning of the week (22 dB). On the contrary, the female broadcaster has an increase in the Harmonics-to-Noise ratio (HNR) in the sustained vowel /a/ recorded at the end of the week (25.5 dB) compared to the beginning of the week (17.6 dB).

Figure 8.

Harmonics-to-Noise ratio of female and male broadcasters during the baseline (pre-monitoring Monday) and follow-up (post-monitoring Friday) measures. The voice production corresponds to the sustained production of the vowel /a/.

Table 1 shows occupational noise levels, Fo, and SPL during broadcasting. In the male broadcaster, Fo, SPL, and occupational noise exposure are higher on Monday compared with Friday. In the case of the female broadcaster, although occupational noise levels during both monitoring (Monday and Friday) were similar, Fo is higher on Monday compared with Friday.

Table 1.

Noise level exposure and occupational voice use of two broadcasters during one-week follow-up

Broadcaster	Weekday	Noise level	Fo	SPL	Duration of the exposure
Male	Monday	96 dBA	161 Hz (47 Hz)	86 dB (12 dB)	1 hour day
	Friday	89 dBA	152 Hz (45 Hz)	79 dB (13 dB)	1 hour day
Female	Monday	75 dBA	171 Hz (56 Hz)	83 dB (14 dB)	4 hour day
	Friday	76 dBA	160 Hz (38 Hz)	83 dB (18 dB)	4 hour day

Occupational noise exposure during broadcasting

As shown in Table 1, noise levels during radio broadcasting were below Occupational Exposure Limits. However, noise levels were higher during broadcasting of the sports show with the male broadcaster than during the morning news with the female broadcaster. Figure 9 and 10 show the noise exposure levels during the total duration of the monitoring of the female and male radio broadcasters.

Figure 9.

Background noise levels (dB(A)) for 4-hours radio shows on Monday and Friday of the female broadcaster

Figure 10.

Background noise levels (dB(A)) for 1-hour radio shows on Monday and Friday of the male broadcaster

Short-term effect of noise exposure during radio broadcasting on hearing function

The evaluation of the effect of radio broadcasting on hearing function was performed by comparing the information collected during the pre-monitoring assessment on Monday (baseline) with the post-monitoring assessment on Friday (follow-up). Neither the female broadcaster nor the male broadcaster reported any current hearing problem (ringing in ears, itchy sensation, pain in ears, difficulty hearing or understanding spoken communication) at the baseline or at the follow-up.

No significant changes in the hearing thresholds were observed in the broadcasters after broadcasting. As shown in Figure 11, while the female broadcaster had a mild decrease in hearing threshold level in the frequency 8000Hz, her pure tone average (500Hz – 1000Hz – 2000Hz, PTA) were under 20 dB, which is considered within normal limits. Figure 12 shows that the male broadcaster results during the one-week follow-up were also within normal limits.

Figure 11.

Audiograms of the female broadcaster before and after broadcasting on Monday and Friday. Upper audiograms correspond to Monday. Lower audiograms correspond to Friday. Left audiograms correspond to pre-monitoring assessment. Right audiograms correspond to post-monitoring assessment. Circle corresponds to right ear. “X” corresponds to left ear.

Figure 12.

Audiograms of the male broadcaster before and after broadcasting on Monday and Friday. Upper audiograms correspond to Monday. Lower audiograms correspond to Friday. Left audiograms correspond to pre-monitoring assessment. Right audiograms correspond to post-monitoring assessment. Circle corresponds to right ear. “X” corresponds to left ear.

DISCUSSION

The aim of this study was to explore the short-term effect of work-related voice use on voice function, and noise exposure on hearing function among radio broadcasters. For the two broadcasters studied, three main results were found. First, changes in fundamental frequency post-monitoring at the end of the work week may be an indication of work-related vocal fatigue. Second, changes in the distribution and standard deviation of SPL during the monitoring on Friday compared with Monday may be an indication of broadcasters´ control of their vocal loudness as a strategy to reduce vocal effort during broadcasting. Third, during a one-week follow-up, noise conditions during radio broadcasting were below the permissible exposure values established by the Occupational Safety and Health Administration (OSHA) (24), and without consequences on hearing function.

Regarding voice production during broadcasting (occupational voice use), one interesting finding was the change on the distribution of the broadcasters´ vocal sound pressure levels on Friday compared with Monday. The distribution of SPL of the female broadcaster changed from Monday to Friday, with a more disperse distribution on Friday. The results also show higher standard deviation on Friday (SPL SD= 18 dB) compared with Monday (SPL SD= 14 dB). The male broadcaster had the highest peak of the distribution is slightly skewed to the right on Monday compared with Friday (figure 6), and slightly higher standard deviation on Friday (SPL SD= 13 dB) compared with Monday (SPL SD= 12 dB). Previous studies have reported that teachers, who are regular voice users, with higher standard deviations were less likely to report voice complaints (6). This change in the distribution of SPL and increase in the standard deviation may be an indication of control of the vocal loudness as a strategy to reduce vocal effort during broadcasting. Nevertheless, future research is needed to confirm these findings.

The results on phonation time percentage (D_t%) showed a high variability between the two participating broadcasters. The mean Dt_% during the follow-up was 39.5% for the male broadcaster, compared to 12.5% for the female broadcaster. The phonation time percentage of the male broadcaster is considerably higher compared with the female broadcaster, and compared with previous studies among call-center operators that reported phonation time percentage of 14.7% (4). It is almost certain that the reason for this higher phonation time percentage of the male broadcaster is that his vocal production is almost entirely a “monologue” speaking style during the one-hour long sports show, compared to the female’s shared broadcast. Therefore, his vocal demands are high not just in terms of vocal quality but also in terms of vocal load. The Dt_% of the female radio broadcaster suggests a moderate work-related vocal demand in terms of vocal load.

As shown in figures 3 and 5, there was a slight decrease in fundamental frequency (Fo) during the occupational voice monitoring on Friday compared with the monitoring on Monday for both broadcasters. This decrease may be an indication of a more hypofunctional occupational voice use towards the end of the week (39), as a strategy to reduce vocal effort during broadcasting. Nevertheless, future studies with bigger sample sizes and higher number of repeated measures are needed to corroborate this hypothesis. The mean SPL during one-week of monitoring was about 83 dB for both broadcasters. However, we can observe a different distribution of SPL between the two broadcasters, especially on Friday. This finding reflects the different demands on voice production that radio broadcasting requires.

Our results on the short-term effect of radio broadcasting on voice production indicate a tendency to slightly increase the fundamental frequency (Fo) in the speech sample (reading) recorded at the end of the week (maximal work-related voice use) compared with the beginning of the work week (pre- work-related voice use). Previous research has suggested that increased Fo after vocal loading tasks may be an indication of vocal fatigue (18, 40–42). From the physiological point of view, it has been hypothesized that when a speaker experiences vocal loading, the thyroarytenoid (TA) becomes fatigued, which inhibits maintaining a lower pitch, and a greater level of fatigue may eventually be the result (40). A second hypothesis is that higher Fo after vocal loading tasks may reflect an increased muscle tonus as an adaptation to loading (43, 44).

Since no previous studies were found on the short-term effect of radio broadcasting on voice production among broadcasters, it is not possible to compare our results. Nevertheless, our findings are in line with previous results reported among other occupational voice users (15, 43). Among teachers, Laukkanen et al (2008) reported a significantly higher fundamental frequency (around 6 Hz) post-monitoring of one working day. Comparable results have also been reported among call-center workers who had significantly higher Fo (around 5 Hz) at the end of the working day compared with the beginning of the day. Therefore, it is likely that although radio broadcasters have lower vocal demand in terms of vocal load compared with other occupational voice users such as teachers or call center workers (2), the high demand in terms of voice quality would result in similar patterns of voice production. These adjustments on the voice production may be linked with the occurrence of work-related vocal fatigue.

In the post-monitoring voice assessment at the end of the week, we found a different tendency between the male and the female broadcasters on the Harmonics-to-Noise ratio (HNR). The male broadcaster has a decrease in the Harmonics-to-Noise ratio (HNR) in the sustained vowel /a/ recorded at the end of the week compared to the beginning of the week. On the contrary, the female broadcaster has an increase in the Harmonics-to-Noise ratio (HNR) in the sustained vowel /a/ recorded at the end of the week compared to the beginning of the week. The results on the relation between HNR and vocal loading are contradictory in the literature. Some studies suggest no changes on HNR after vocal loading tasks (45), whereas other studies report a decrease in HNR after 1 hour of constant loud reading (41). Futures studies are needed to clarify this relation.

An interesting result in this study was the low values for Fo for the female broadcaster (171 Hz on Monday, and 160 Hz on Friday). Previous research have reported that broadcasters, particularly females, use a low-pitched voice in broadcasting to convey authority and confidence (46, 47). It has also been reported that too high pitch during broadcasting compared with speaker´s optimal pitch may result in tension. Therefore, one possible explanation is that the female broadcaster produces a lower-pitch voice during her show, having naturally a low-pitch conversational voice. It seems likely that, among radio broadcasters, the slightly increased Fo at the end of the working week during the follow-up may be an indication of vocal fatigue due to the occupational voice use. Future studies with a longer follow-up time are suggested in order to explore the natural course of vocal fatigue associated with radio broadcasting.

Our results on the work-related noise exposure during broadcasting indicate that noise levels during radio broadcasting were below the national and international standards. Moreover, no significant changes in the hearing thresholds were identified. Comparing these results with previous research of other occupational voice users, we found that our results are in line with studies that included call-center workers (48, 49). For example, Trompette et al (2012) found that 110 out of 117 call-center workers were exposed to noise levels (Leq,8h) under 80dB(A) considering minimal risk for hearing problems. Similar results were reported by Patel et al (2002) who found that the noise exposure of 150 call-center workers were unlikely to exceed the 85 dB(A) and, therefore, resulted in a minimal risk of hearing problems. Nevertheless, it is important to emphasize that constant, even minimal, noise exposure may still have long-term consequences (50). Exposure to moderate-loud noise can result in a temporary and/or permanent threshold shift. While a temporary threshold shift may fully recover within 24 to 48 hours (51), a previous study on mice found that temporary threshold shift may be associated with an increased nerve degeneration and accelerated age-related hearing loss (52). Further, permanent threshold shift is irreversible. Therefore, noise exposure consequences on hearing mechanisms may occur after years of moderate to loud noise exposure. In this study, although the participating broadcasters were occupationally exposed to noise levels below the standards, and they did not present a shift on their hearing thresholds, the background noise levels during broadcasting were higher than what is recommended for environments where communication is important (53), and also higher than the change-point for self-reported discomfort reported in previous research (54). For this reason, we recommend a conservative approach, and performing an annual hearing assessment among these radio broadcasters in order to monitor their hearing function and prevent future hearing problems.

On the other hand, occupational voice use under high noisy conditions may lead to overuse or misuse of the voice, and therefore to develop voice disorders (2). In the presence of noise, the voice is masked, and its production should be modified to guarantee the success of the communication process. The vocal response by a talker to the background noise conditions of a space is called Lombard effect (16, 55). Among occupational voice users, a Swedish study in 10 daycare centers found that working under mean background noise levels of 76 dB(A) caused teachers to speak on average 9.1 dB louder and with higher mean Fo during work as compared to the baseline (56). Therefore, although our results suggest that noise levels during radio broadcasting were below the national and international standards, occupational voice use under these background noise conditions may cause higher vocal effort, and as consequence vocal fatigue or other voice disorders may appear. Future studies with bigger sample sizes are needed to confirm our results.

The main limitation of this study was the very small sample size that prevents the generalization of our results to the population of radio broadcasters, as well as types of broadcasts (e.g., news, sports, radio), characteristics of broadcasters (gender, age) specifically, and effect of individual intra-variability on the changes of Fo and vocal SPL. For this reason, future research is planned to increase the sample size in order to further explore these issues. Additionally, while it appears broadcasters in this preliminary study may not be in highly noise-exposed groups, it is nevertheless important to continue an examination of this issue. Additional studies should also be designed to explore the use of headphones during radio broadcasting (and potentially other occupations). Its use may increase the risk to be exposed to “acoustic shock” (defined as the sudden increase in the noise transmitted by the headphones due to different causes (48)), which can lead to symptoms like ear pain, tinnitus, and headache, usually without hearing loss (57). Previous research has studied acoustic shock among call-center workers (48, 57) but no studies were found for radio broadcasters. In addition, since voice production is a multifactorial process, it is likely that the reported changes on voice parameters may be also associated with other individual factors, such as “natural” voice changes during the day. Future research is advisable to define the changes on voice parameters during a day without occupational voice use among broadcasters. Another limitation is the lack of information on the participants’ non-occupational voice use. Since, participating broadcasters did not register their non-occupational voice use during the week of the monitoring, it is difficult to define in which proportion voice and hearing function changes were consequence of the work-related exposure. Nevertheless, to the best of the authors knowledge, participating broadcasters did not have high vocal load on non-occupational settings during the week of the monitoring.

CONCLUSION

The work-related communicative profile of radio broadcasting indicates that although vocal demands in terms of vocal load may differ among broadcasters, the work-related conditions of broadcasting may play a role in vocal function among these occupational voice users. Concerning the hearing function, our results indicate that occupational noise exposure be represented minimal risk for hearing problems but the consequences of long-term noise exposure on hearing mechanisms may yet occur. Future studies with bigger sample sizes are warranted to confirm our results.