|Year : 2016 | Volume
| Issue : 1 | Page : 7-13
Singer's formant in Hindustani classical singers
Namita Joshi Associate Prof. 1, Manna A Raju2
1 Department of Speech Language Pathology, School of Audiology and Speech Language Pathology, Pune, Maharashtra, India
2 MASLP, Kochiyil Poovakala Bethel, Pathanamthitta, Kerala, India
|Date of Web Publication||5-Apr-2017|
School of Audiology and Speech Language Pathology, Pune - 410 043, Maharashtra
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Context: Since the last two decades, the research has been concentrated on professional voice; there is a curiosity among researchers to find out whether singers (S) are really special in their voice use than nonsingers (NS). Although there have been studies regarding singer's formant in Indian classical singers, underlying physiological adjustments made by singers and its relation with acoustical features need to be studied. Thus, the aims of the present study were to explore the possibility of singer's formant in Hindustani classical singers, to study the physiological adjustments made by the singers, and also to find out the effect of number years of experience on acoustical as well as auditory-visual perceptual parameters. Subjects and Methods: Twenty Hindustani classical singers and ten nonsingers within the age range of 18–50 years were considered for the study. The singers were again grouped into three Subgroups (S1, S2, S3) based on their years of music experience. Acoustical as well as auditory-visual perceptual evaluation was carried out for singers and nonsingers. Results and Discussion: The study revealed that S3 (>10 years of experience) Group had the highest amplitude at 2–4 kHz as well as higher singing power ratio values as compared to S1 and S2. Nasoendoscopic evaluation was done for all the participants by a well-experienced ENT professional. S3 Group had the widest pharyngeal opening. Hence, the results of the present study point out the possibility of occurrence of singer's formant in Hindustani classical singers. Conclusions: The result of the acoustical analysis indicates that there is an effect of years of music experience on the acoustical as well as perceptual parameters. The study highlights the possibility of developing training material for singers to improve their pharyngeal widening.
Keywords: Acoustical and perceptual analysis, nasoendoscopy, Singer's formant, singing power ratio
|How to cite this article:|
Joshi N, Raju MA. Singer's formant in Hindustani classical singers. J Laryngol Voice 2016;6:7-13
| Introduction|| |
Singing is defined as “sensory motor phenomenon that requires particular balanced physical skill.” Maintenance of a better quality and wide range of voice for a longer period is a necessity for singers (S). Awareness about voice culture is increasing among singers and they often approach professionals (Speech Language Pathologist and Otolaryngologist) for care and maintenance of singing voice. These professionals play a crucial role in the assessment and management of singing and speaking voice. Singers are special when compared to nonsingers (NS) due to an exciting “ringing” voice quality during a singing performance that helps them to be heard over a large orchestral accompaniment., The author explained about “brightness in voice” as “singer's formant” which occurs as a result of clustering of higher third, fourth, and fifth vowel formants (F3, F4, and F5). To find out whether singers are special in their vocal usage, various authors studied following long-term average spectrum (LTAS) parameters: amplitude at lower harmonics (0–2 kHz), amplitude at higher harmonics (2–4 kHz), center frequency of singer's (FS) formant, and singing power ratio (SPR).,,,,,
However, research on the vibratory mechanism of vocal folds and its relation to acoustical parameters are rare. Examination of the vibratory mechanism of vocal folds can be done using various instruments such as videofluoroscopy, nasoendoscopy, videolaryngoscopy, and strobovideolaryngoscopy. Researchers have reported an association between pharyngeal widening and production of ringing voice while singing. Singers with widened pharynx were observed to have effective clustering of higher formants and thereby higher singers' formant. Perceptual evaluation has been considered as a gold standard for assessing voice though the subjectivity can be reduced by complementing it with acoustical evaluation. Singing is a complex act and its uniqueness can be investigated by a combination of auditory-visual perception as well as acoustical analysis.
In the Western scenario, research concentrated on finding out singer's formant, center frequency, and SPR in singers and nonsingers. FS is defined as a pronounced peak of acoustic energy which is located around 3 kHz and it is formed by the clustering of the third, fourth, and the fifth vowel formants (F3, F4, and F5). Author has stated that the center frequency (frequency at which the highest peak amplitude between 2 and 4 kHz is obtained) of singer's formant lies between 2.4 and 3.6 kHz in almost all of the singers. In Indian classical singers, there have been attempts to find out the center frequency of FS. It was found that maximum energy concentration observed to be around 2500–2950 Hz in trained Carnatic singers and around 2300–2500 Hz in untrained singers. FS was located in 2–4 kHz region in Hindustani classical singers and its center frequency increased with the rise in pitch, which was parallel to sung vowels in Western music. It has been found that singer's formant is present in Indian classical singers. Contrary to the above-mentioned studies, singer's formant was not found in South Indian classical singers. Although there have been studies regarding singer's formant in Indian classical singers, underlying physiological adjustments made by singers and its relation with acoustical features need to be studied. The review of literature suggests that there have been very limited number of studies done to explore the physiological differences in singers and nonsingers during singing. There have also been limited researches on the effect of years of experience on the acoustical and perceptual parameters. Hence, the present study was planned to explore the physiological, acoustical, and perceptual differences between singers (with varied experience) and nonsingers. The aim of the study was to explore the presence of singer's formant in Hindustani classical singers, to investigate physiological adjustments made by the singers and to find out the level of impact of a number of years of experience of singing on singer's formant.
| Subjects and Methods|| |
A total of 20 Hindustani classical singers (11 males and 9 females) and 10 nonsingers (5 males and 5 females) were participated in the study. The mean age of singers was 29.15 years and that of nonsingers was 22.5 years. The Hindustani classical singers were again divided into three Subgroups S1, S2, and S3 according to their years of experience in singing. S1 included those singers who had a formal vocal training of about 0–5 years (4 males and 3 females). S2 singers had an experience of about 6–10 years (4 males and 3 females). Those singers with more than 10 years of experience were considered in Subgroup S3 (3 males and 3 females). Nonsingers were those participants who had not undergone any kind of formal vocal training. Case history revealed that singers in S1 Group used to practice around an average of 2–3 h/day. Most of them were students of the college of fine arts where they had just started formal training, whereas singers in Group S2, have a practice of singing for 4–5 h/day. Many of them have an experience of singing in semi-classical as well as Indian classical music. Singers in Group S3 were the most experienced singers with maximum knowledge about voice culture. They have a practice of more than 6–7 h/day. They were involved in some of the voice care practices such as drinking sips of water during performances, steam inhalation, and Omkar practice before daily singing activity (Riyaz). All the participants were included only when they did not show any sign of any speech, language, vocal, sensory motor problems at the time of recording. Before the recording, all the participants were explained regarding the purpose of the study and the future implications. Consent form was filled by all the participants.
The stimuli recorded for the particular study was a phonation sample of the vowel /a/, /i/, and /u/ for about 5 s and also the well-known song/vanðe mataram/for about 20 s. The acoustical analysis was carried out on the recorded stimuli to see for the presence of singer's formant. A portable digital tape recorder (SONY, ICD/VX/533F) was used for recording the tasks. The microphone was kept away from the participant at a distance of 10 cm. All recordings were done in a quiet environment with the participants seated in comfortable position.
The recordings of a sample of song were transferred to the computer, and acoustical analysis was done using LTAS of CSL 4500 MDVP software at a sampling rate of 44,000 Hz.
Long-term average spectrum parameters
- Amplitude of the highest peak at the lower harmonic (0–2) kHz region and the amplitude of the highest peak at the higher harmonics (2–4) kHz region
- Mean center frequency of singer's formant: It is the peak frequency of the highest amplitude obtained in region of higher harmonics (2–4 kHz)
- SPR: It is the ratio of amplitude obtained at 2–4 kHz to that of 0–2 kHz. SPR is a parameter that can quantitatively measure the resonant quality of singing voice 
- Mean fundamental frequency (mean F0): It is the average frequency in a periodic waveform.
Auditory-visual perceptual analysis
Nasoendoscopy was carried out to investigate the physiological adjustments made by the singers with respect to the presence of singer's formant when compared to nonsingers. A 70° rigid nasoendoscope (HAWK model) with a diameter of 4 mm was used for the analysis. It was kept at 7–8 cm from the opening of the nostrils of the anterior nares. The task of the participant was to phonate /i/ while nasoendoscopy will be done simultaneously with the help of an experienced ENT professional. The videos were saved on a compact disk drive.
Auditory-visual perceptual evaluation of velopharyngeal closure pattern, lateral and posterior wall movement, and also the pharyngeal widening was carried out. Four well-experienced ENT professionals were asked to rate the nasoendoscopy video samples. All the samples were randomly presented to the raters. The inter-rater reliability was estimated to see the correlation between the visual perceptual evaluation. Parameters measured for auditory-visual perceptual evaluation are as follows.
Velopharyngeal closure pattern and lateral or pharyngeal wall movement
The professionals were asked to rate nasoendoscopic videos on the basis of two parameters. One is velopharyngeal closure pattern: coronal, sagittal, circular and second is the lateral versus pharyngeal wall movement.
The ENT professionals were asked to rate the nasoendoscopy samples on a three-point rating scale. (1 = Narrow pharyngeal opening, 2 = Medium pharyngeal opening, 3 = Wider pharyngeal opening).
Results of acoustical and auditory-visual perceptual evaluation are tabulated and discussed.
| Results and Discussion|| |
Recorded sample of all the participants was transferred to CSL-MDVP module for LTAS analysis. The results of acoustical and auditory-visual perceptual analysis are as follows:
Long-term average spectrum parameters
Amplitude at 0–2 and 2–4 kHz
The amplitude of the highest peak at the lower harmonics (0–2) kHz region and the amplitude of the highest peak at the higher harmonics (2–4) kHz region were analyzed.
[Figure 1] shows that there was no much difference in amplitude at 0–2 kHz for nonsingers and singers as a group. However, the amplitude at 2–4 kHz was found to be higher in singers than for nonsingers. When the comparisons within the group of singers were considered, it was found that amplitude at 0–2 kHz remained almost similar for all the three groups of singers. The amplitude at 2–4 kHz was found to be variable for singers. Higher amplitude was noticed for singers with more than 10 years of experience (S3) when compared to singers of 0–5 (S1) and 6–10 (S2) years of experience. Although the difference was observed, it was not statistically significant. The participants in the Groups S1 and S2 were students of the musical institute, and the amount of singing and daily Riyaz carried out by them was not as much as Group S3. The variability in amplitude at 2–4 kHz may be due to the difference in the number of hours of the Riyaz they were doing per day. The findings of the study showed that there is a tendency for the increment in amplitude level at 2–4 kHz as the years of music experience increase. Thus, the amount of training and hours of practice per day can be expected to be a supplementary factor for the achievement of significant energy concentration at higher harmonics. The findings of the present study support the observation of previous studies, where they have found that as the years of experience or the vocal training increases, amplitude (energy levels) too increases. Authors also observed additional energy concentration in 2–4 kHz region for trained singers when compared to nonsingers., The higher amplitude obtained at 2–4 kHz region for singers with more than 10 years of experience (S3) can be correlated to pharyngeal widening (supralaryngeal adjustments) observed in the auditory-visual perceptual analysis for S3 Group. The results of the present study can be confirmed on a larger sample size. It can also be assumed that rigorous singing training for less than 10 years may be a too short time period to expect an effect in the higher harmonic region.
|Figure 1: Mean scores of amplitude at 0–2 and 2–4 kHz for singers and nonsingers|
Click here to view
Mean center frequency of singer's formant
The center frequency of FS was analyzed to see the frequency at which energy is concentrated.
In the present study, the maximum energy concentration in singers was observed from 2.1 to 3.4 kHz with the mean center frequency of 2.4 kHz, 2.9 kHz, and 2.7 kHz for S1, S2, and S3, respectively. Although the differences between mean center frequencies for the groups were obtained, it was not statistically significant. The above findings were consistent with the prior investigations, where it was found that for sopranos and mezzo-sopranos, a significant energy concentration is observed in the higher harmonics region which is in 2.8–3.4 kHz region (P < 0.05). Singer's formant was observed to be near 3 kHz for a well classically trained bass, baritone, tenor, and alto singers' voices. The present study did not consider the different voice categories in Hindustani classical singers (e.g., Drupad, Khyal, Tumri) like in Western studies mentioned above. The group of singers was more heterogeneous and small to generalize the results for a range of center frequency in Hindustani classical singers.
Singing power ratio
It is calculated by the ratio of the highest peak intensity level obtained at higher harmonics (2–4) kHz to the highest peak intensity at lower harmonics (0–2) kHz. As shown in [Figure 2], among the three groups of singers, S3 (0.6) demonstrated the highest SPR value than S1 (0.5) and S2 (0.5). In the present study, a tendency of an augmentation in the SPR values was observed as the music training increases. Although the differences were seen for the groups, these differences were not statistically significant. The results of the present study were similar to that of studies done in the past, where they found that professional singers with more years of music experience (>4 years) had significantly higher SPR values than trained singers with less than 4 years of experience. The results of the present study were unlike the studies  where they found that SPR scores were significantly lesser for singer group than for nonsinger group. Higher SPR values in S3 Group show that they had higher energy concentration at 2–4 kHz than 0–2 kHz, which confirms the possibility of formant clustering at higher harmonics. The expertise and self-regulating pharyngeal widening in S3 Group would have led them to have higher SPR values as compared to S2 and S1. The present study highlights the possibility of obtaining the singer's formant for singers with more than 10 years of experience in the region of 2–4 kHz.
|Figure 2: Mean value scores of singing power ratio of singers of each group and nonsingers|
Click here to view
Mean fundamental frequency (mean F0)
The fundamental frequencies of the participants were analyzed acoustically. The gender-wise comparison was done for the parameter of mean F0 between singers and nonsingers. Independent sample t-test was carried out which revealed that there is no significant difference in mean F0 between males and females within the group and across the group [Figure 3] and [Figure 4]. The results of the present study show that singers were not special than nonsingers in terms of their mean F0. The mean F0 values for female singers were higher in comparison with female nonsingers which supports the other studies, where they also found that the mean relative F0 of trained females (soprano, mezzo-soprano, and altos) was higher than female nonsingers. However, the results were not similar for male category, where the male singers were found to have lower F0 values than male nonsingers. This probably is the result of low magnitude of jaw opening in male nonsingers during the task of vowel phonation as also reported. The differences obtained in the present study can be due to the less sample size and unequal distribution of males and females across the groups.
Auditory-visual perceptual parameters
Velopharyngeal closure pattern and lateral/posterior pharyngeal wall movement are the auditory-visual perceptual parameters.
The velopharyngeal closure pattern of all the participants was perceptually evaluated with the help of four experienced ENT professionals in Pune. The percentage of occurrence of velopharyngeal closure patterns in each of the subgroups of singers and that of nonsingers was calculated. [Table 1] shows the percentage of velopharyngeal closure pattern in singers and nonsingers. For nonsingers, coronal pattern was observed in 60% of the participants, circular pattern was observed in 40%, and none of them had sagittal pattern. For the subgroup of singers in S1, 57.14% of the participants had coronal pattern, 28. 57% had sagittal pattern, and 14.28% had circular pattern. For Subgroup S2, coronal pattern was present for 57.14% of participants, sagittal were present for 14.28%, and circular was present for 28.57% of the participants. Around 66.66% of the participants of Subgroup S3 singers had coronal pattern with remaining 16.66% of participants with sagittal as well as circular pattern. As shown in [Table 2] (overall lateral or posterior pharyngeal wall movement in singers and nonsingers), all the groups had lateral wall movement. The most common velopharyngeal closure pattern was coronal with lateral wall movement across the groups. Hence, the results of the present study conclude that singers were not special in the velopharyngeal closure pattern than nonsingers.
|Table 1: The percentage of velopharyngeal closure pattern in singers and nonsingers|
Click here to view
|Table 2: Overall lateral or posterior pharyngeal wall movement in singers and nonsingers|
Click here to view
Studies are limited regarding the exploration of velopharyngeal closure pattern as well as lateral or posterior pharyngeal wall movement in singers and nonsingers.
[Table 3] shows an overall rating of pharyngeal widening for singers and nonsingers. The groups of NS, S1, and S2 were rated 2 for pharyngeal widening, which means they had a medium pharyngeal opening. S3 Group was rated 3, which means they had a wider pharyngeal opening.
|Table 3: Overall rating of pharyngeal widening for singers and nonsingers|
Click here to view
The product moment correlation between each of the raters for pharyngeal widening is significant (P < 0.05).
Singer's formant cluster is created in the vocal tract by narrow epilaryngeal tube opening so that the ratio of the cross-sectional area of the pharynx is approximately six times greater than that of epilarynx outlet (1:6 ratio). The maintenance of an enlarged pharyngeal area is essential for the achievement of pharyngeal – laryngeal vestibule area ratio (6:1). The sample of nasoendoscopy evaluation of singers and nonsingers was evaluated by four experienced ENT professionals on a three-point rating scale (1 = Narrow pharyngeal opening, 2 = Medium pharyngeal opening and 3 = Wider pharyngeal opening). Inter-rater reliability was carried out for the parameter of pharyngeal widening [Table 4].
|Table 4: Inter-rater reliability for pharyngeal widening between singers and nonsingers|
Click here to view
It can be opined that well-trained singers use more efficient supralaryngeal adjustments which enables them to produce a bright vocal quality. The results of auditory-visual perceptual evaluation of pharyngeal widening are in support of acoustical findings of higher SPR values in Group S3. The present study postulates the fact that more the number of years of experience, better are the supralaryngeal adjustments and thereby higher energy concentration in the FS region. The result of the present study is parallel to the results obtained by other authors, where they revealed that trained singers (8–10 years of experience in Carnatic singing) showed clustering of higher formants due to widened pharynx.
The results of the study may be contributed to the fact that the singers with <10 years of experience are still undergoing formal music training and have not achieved mastery in singing as they are no difference in terms of acoustical and auditory-visual perceptual parameters with that of nonsingers. Hence, singer's formant can be considered as a learned behavior which could be attained only when one masters in making physiological adjustments with more than 10 years of singing experience. Thus, pharyngeal widening should be the part of training for professional singers who need to increase the projection of their voice.
The results of the present study indicate the existence of singer's formant in Hindustani classical singers. The results of the acoustical analysis point out the effect of years of singing experience on all the parameters (LTAS parameters and F0). The S3 Group had the highest amplitude at 2–4 kHz, had the highest SPR values as compared to S1, S2, and implies the formant clustering in higher harmonics. According to auditory-visual perceptual evaluation, S3 Group had the widest pharyngeal opening which is in correlation with the results of the acoustical analysis. Hence, the study concludes that singers with more than 10 years of experience are special in terms of acoustical and auditory-visual perceptual parameters of voice than nonsingers though the significant difference between the groups could not be found due to the small sample size and unequal distribution of participants within the group. The present study suggests that for improving the voice quality in professional voice users, proper physiological adjustments (pharyngeal widening) should be taught. The study highlights the possibility of developing training material for singers to improve their pharyngeal widening. The study points out the need and development of an objective parameter (SPR) for differentiating style of singing in future. It throws light into the future utility of FS in pathological conditions among Hindustani classical singers. The study can also be repeated to compare SPR values and nasoendoscopic evaluation between different styles of singing such as Carnatic versus Hindustani classical singers.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Bunch M. Dynamics of the Singing Voice. New York: Springer, Verlag; 1982.
Sundberg J. The Science of the Singing Voice. Dekalb, Illinois: Northern Illinois University Press; 1987. p. 35.
Sundberg J. What's so special about singer's?. J Voice 1990;4:107-19.
Barrichelo VM, Heuer, RJ, Dean CM, Sataloff RT. Comparison of singer's formant, speaker's ring, and LTA spectrum among classical singers and untrained normal speakers. J Voice 2001;15:344-50.
Ghosh K. Comparison of Acoustic Characteristics in Female Trained (Carnatic Style) Singers, Untrained Singers and Non Singers. Unpublished Master's Dissertation. University of Mysore; 2007.
Lundy DS, Roy S, Casiano RR, Xue JW, Evans J. Acoustic analysis of the singing and speaking voice in singing students. J Voice 2000;14:490-3.
Santhosh M, Devie N, Yeshoda K. Acoustical Analysis of Voice of Singing Students. Proceedings of the 35th
Indian Speech and Hearing Association Conference; 2008.
Sheetal R, Yeshoda K. Acoustical analysis of singer's voice. JISHA 2008;22:102-5.
Sundberg J, Birch P, Gümoes B, Stavad H, Prytz S, Karle A. Experimental findings on the nasal tract resonator in singing. J Voice 2007;21:127-37.
Sundberg J. Articulatory interpretation of the “singing formant”. J Acoust Soc Am 1974;55:838-44.
Sundberg J. Level and center frequency of the singer's formant. J Voice 2001;15:176-86.
Sujatha S. Analysis of Singing Voice. Unpublished Master's Dissertation. University of Mysore; 1989.
Sengupta R. Study on some aspects of the singers formant in North Indian classical singing. J Voice 1990;4:129-34.
Devie N. Singer's Formant in Indian Classical Singers. Carnatic vs. Hindustani. Unpublished Master's Dissertation. University of Mysore; 2003.
Boominathan P. Do Carnatic Singer's Have Singer's Formant? – An Exploratory Study. Proceedings of The Frontiers of Research in Speech and Music Conference; 2004.
Omori K, Kacker A, Carroll LM, Riley WD, Blaugrund SM. Singing power ratio: Quantitative evaluation of singing voice quality. J Voice 1996;10:228-35.
Smits I, Ceuppens P, De Bodt MS. A comparative study of acoustic voice measurements by means of Dr. Speech and Computerized Speech Lab. J Voice 2005;19:187-96.
Watts C, Barnes-Burroughs K, Estis J, Blanton D. The singing power ratio as an objective measure of singing voice quality in untrained talented and nontalented singers. J Voice 2006;20:82-8.
Oliveira Barrichelo VM, Heuer RJ, Dean CM, Sataloff RT. Comparison of singer's formant, speaker's ring, and LTA spectrum among classical singers and untrained normal speakers. J Voice 2001;15:344-50.
Reid KL, Davis P, Oates J, Cabrera D, Ternström S, Black M, et al.
The acoustic characteristics of professional opera singers performing in chorus versus solo mode. J Voice 2007;21:35-45.
Lee SH, Kwon HJ, Choi HJ, Lee NH, Lee SJ, Jin SM. The singer's formant and speaker's ring resonance: A long-term average spectrum analysis. Clin Exp Otorhinolaryngol 2008;1:92-6.
Akerlund L, Gramming P, Sundberg J. Phonetogram and averages of sound pressure levels and fundamental frequencies of speech: Comparison between female singers and nonsingers. J Voice 1992;6:55-63.
Lim M, Lin E, Bones P. Vowel effect on glottal parameters and the magnitude of jaw opening. J Voice 2006;20:46-54.
Monson BB. The 1:6 ratio in vocal pedagogy. J Acoust Soc Am 2007;3:1728-36.
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2], [Table 3], [Table 4]