Participants
A total of 50 subjects (34 females and 16 males) aged 18–39 years participated in this study. All subjects were right-handed students of the Rehabilitation School (Beheshti University of Medical Sciences, BUMS, Tehran, Iran) with normal hearing, which was confirmed by normal results from otoscopy, the immittance test (tympanometry and acoustic reflex), and pure tone and speech audiometry. None of the participants had any history of neurologic pathologies. All participants signed a written informed consent. All methods were carried out in accordance with relevant guidelines and regulations of BUMS. All experimental protocols were approved by Human Research Ethics Committee (BUMS, Ethical Code No. IR.SBMU.RETECH.REC.1396.571).
Music Selection
In the first phase of the study—selecting songs—we chose 10 songs out of the 25 most popular songs from bestselling Persian song albums (in the reference of Keihan Newspaper, Tehran, Iran), as well as those with the highest number of downloads (https://www.radiojavan.com/) in the last two years. All songs, originally Persian, were played for 50 people who were then asked to subjectively rate each song from 1 (low) to 10 (high). Based on these ratings, 10 songs were finally chosen.
In the second phase of the study—ranking songs—all participants were asked to listen to the ten selected songs as much time as they wanted to. Then, they were asked to rate them on a visual scale from 1 (less preferred) to 10 (highly preferred). The score was recorded separately for each song in a form, all song-related components were addressed—the voice of the singer, rhythm, and subjective feelings post listening (Table 1).
In the third phase, the songs were saved as separate “wav.” files in a computer, and the root mean square (RMS) of all the songs was derived using Adobe Audition software (Adobe Co, 2017, USA).
Table 1. Music tracks used as pleasant and unpleasant music in terms of the means of subjective scores provided by participants
Song & Singer
|
Mean Score
|
30 Salegi
by Ehsan Khajeh Amiri
|
7.84
|
Roya-ye Bi Tekrar
by Ali Zand Vakili
|
7.02
|
Harmless Ruler
by Mohsen Chavoshi
|
6.70
|
The Road’s Dancing
by Charttaar band
|
6.14
|
Full Length Mirror
by Mehdi Yarrahi
|
5.76
|
Dele Majnoon
by Mohammad Reza Shajarian
|
5.18
|
Khoda Hamin Havalie
by Hamed Homayoun
|
5.02
|
Ta Nafas Hast
by Shahram Shokoohi
|
4.40
|
Ansolute Nothingness
by Hafez Nazeri
|
4.00
|
Manshour
by Kave Yaghmaei
|
2.98
|
The ANL Testing
In sum, the ANL tests were approached in the following ways:
- The typical ANL test, which plays a female speaker’s running speech (as the speech signal) with a 12-talkers babble noise (as the noise signal), and
- The modified ANL test includes composites of various conditions such as a female speaker’s running speech, and less preferred music and highly preferred music (as the speech signal and noise signals, respectively).
For typical ANL testing, the Persian version of the ANL test was used (Tehran University of Medical Sciences, TUMS) [43]. For the modified ANL conditions, the RMS of all songs were calibrated in terms of female speakers’ speech or 12-talker babble noise using Adobe Audition software (Adobe Co, 2017, USA).
All signals were played through headphones (TDH-39) using a laptop (Dell Co, USA), connected via a 2.5 mm audio jack to a clinical audiometer (AC40, Interacoustics Co, Denmark), which was calibrated in concordance with the American National Standards Institute code [44]. Both the volume of the laptop and auxiliary audiometer input were set at 0 volume units using a calibrated tone of 1000 Hz, which was included in the typical ANL Persian test. All the above ANL tests were performed under monotic and dichotic listening conditions. For the monotic listening condition, the right ear was tested; both the signals and noises were presented monoaurally. For the dichotic condition, the signals and noises were presented to the right and left ears, respectively.
Each ANL testing consists of three stages, as listed below:
1) The most comfort level (MCL) measurement: A female speaker’s running speech is presented by a calibrated audiometer at 30 dB HL through a headphone, which the subject is asked to listen and provide feedback about its level. The level of the speech is then increased and decreased, depending on the subject’s signal (thumb up and thumb down, respectively), in steps of 5 dB. Near the final adjustment, 2 dB steps are used for the exact determination of the MCL. After two repetitions, the average is recorded as the MCL measurement.
2) The background noise level (BNL) measurement: While presenting the female speaker’s running speech at the measured MCL, a background noise (12-talker noise) is presented at a starting level of 30 dB HL. The level of noise is increased in steps of 5 dB, and the subject is asked to moderate the level until it becomes intolerable. Again, after reaching the final level of noise, the 5 dB steps are replaced by 2 dB steps. The BNL is the highest level of noise that a subject cannot tolerate. After two repeated measurements, the average is recorded as the BNL. Since there are various vocal and musical components within song, the BNL measurements are obtained in three different and sufficiently lengthy parts of the running music. The averaged measure is then recorded as the BNL for music.
For the dichotic condition, while the running speech or music signals were presented to the right ear at the MCL, babble or music noise as the background noise was presented to the left ear.
3) The ANL measure: The ANL is obtained by subtracting the BNL from the MCL (ANL = MCL - BNL).
In total, ANL measures were obtained for seven conditions under both monotic and dichotic listening. The total duration for testing each participant was approximately 90 min. For each monotic and dichotic listening condition, various other conditions were tested randomly for every subject, and several resting periods were provided.
Statistical Method
SPSS (v. 24.0) was used to analyze the study data (IBM Corp, Armonk, New York, USA). Descriptive statistical parameters, such as the means, standard deviations, and ranges of the MCL, BNL, and ANL results were considered. The Shapiro-Wilk test showed a normal distribution of data. A repeated measures ANOVA test was used. As the overall results were significant, Bonferroni correction was used to determine statistically significant differences for ANLs under different conditions.