|Year : 2013 | Volume
| Issue : 2 | Page : 41-45
Acoustical and perceptual analysis of voice projection in Marathi speaking actors
A Namita Joshi, S Bhushan Jirgale
Department of Speech Language Pathology, Bharati Vidyapeeth University's School of Audiology and Speech Language Pathology, Dhankawadi, Pune, Maharashtra, India
|Date of Web Publication||7-May-2014|
A Namita Joshi
Bharati Vidyapeeth University's School of Audiology and Speech Language Pathology, Dhankawadi, Pune, Maharashtra
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Context: Exploring the characteristics of actor's voice is been the interest of the researchers in last two decades. Research on professional voice users in India is mostly concentrated on singers, teachers, radio jockeys etc. A literature search indicates two areas where more research is needed. Primarily detailing the way of perceptual evaluation of actor's voice (speaking voice) and correlating it acoustic parameters. Research needs to be carried out to find out whether the actors are special as in the strategies they use to project their voices in Indian contexts. Aims: The primary aim of the following study is to analyze acoustic and perceptual parameters for voice projection in actors and non-actors at three loudness levels and secondary aim is to find out correlation between them. Subjects and Methods: A total of 18 actors and 18 non-actors within the age range of 19-60 years participated in the study. Their voices were digitally recorded and analyzed using LTAS (Dr. Speech, Real Analysis-Tiger Electronics, Version 5) on frequency-intensity parameters. Result and Discussion: The result showed that there is a significant difference between actors and non-actors at stage performance level (P < 0.01). Pattern of spectral slope was found sharply falling with higher energy concentration in actors at 3-4 kHz. Perceptual parameters correlated moderately only with average f0 at stage performance level. Conclusion: Actors are efficient in projecting their voice perceptually than non-actors which is in correlation with acoustic parameters at stage performance level.
Keywords: Acoustical and perceptual analysis, actors, long-term average spectrum, vocal projections
|How to cite this article:|
Joshi A N, Jirgale S B. Acoustical and perceptual analysis of voice projection in Marathi speaking actors. J Laryngol Voice 2013;3:41-5
| Introduction|| |
Since ancient days, the actor's performance, vocal needs and vocal hygiene have been studied scientifically, in "Natyaveda", the fifth Veda created by Lord Bramha. The Natyashastra advises about not only what is to be portrayed in a drama, but also how the portrayal is to be executed. The Natyaveda gives guidelines for every performer about how the voice needs to be projected, the emotions to be expressed, body movements to be carried out during the act etc., Actors, considered to be Elite Vocal Performers-Level I, are defined as "persons for whom even a slight aberration of voice may have dire consequence."  For actors (theater artists), effective vocal projection is essential to make it possible for their voices to be heard by listeners with maximum intelligibility with minimum vocal effort. Projection of voice can be defined as the strength of speaking or singing whereby the voice is used loudly and clearly with effective use of laryngeal mechanism. Therefore, understanding of the actor's voice, perception of the acting voice and its role in both performance and clinical assessment is important.  The actor's voice is a product of his/her individual experiences, emotional training and perceptions. The vocal performer learns in large part, by means of teacher's perceptions and his or her perception of the teacher's instructions and feedback. As elite professionals, actors are already blessed with trained ears, to perceive the slightest change in their vocal performance. Hence perceptual assessment of the actor's voice need more formalized scheme. The aim of perceptual analysis should be to evaluate much more (e.g. projections; emotions, resonance quality, diction etc.) than just pathological symptoms in the voice carried out in routine evaluation using rating scales such as grade, roughness, breathiness, asthenia, strain[ 3] and consensus auditory perceptual evaluation-voice.  Usually voice quality is evaluated perceptually, with the parameters which generate certain disagreement among the listeners due to subjectivity involved in this process. Hence perceptual analysis should be complemented with acoustical analysis. Literature shows that researchers ,, have used long-term average spectrum (LTAS) most frequently to bring objectivity. The LTAS is a particularly useful method because it allows the study of persistent long-term voice factors such as projection and quality. Using LTAS  studied male classical singers and formed the concept of "singer's formant" (SF). The SF is a peak at 3 kHz region resulting from the grouping of energy of third, fourth and fifth formants. SF is related to the perception of vocal projection. The study by Nawka et al.  identified in the voice spectrum envelop of Germen actors, an energy increased between 3.1 and 3.7 Hz that would be related to a "loud and shiny" quality of voice and a more gentle fall of actors' voice spectral curve in habitual and strong intensities when compared with normal and moderately rough voices. In the speaking voice  relates a peak at 3.5 kHz, the speaker's formant or actor's formant  studied Australian actor's voices and could distinguish a comfortable projected voice from maximally projected voice. They observed the increased acoustic energy in higher frequency region compared with lower one in LTAS samples with maximal projected voice. Studies have also agreed that correlation between acoustical and perceptual evaluation is important to find out the possible physiological characteristics for betterment of voice quality.
Literature has indicated the need of three important research areas in the field of elite vocal performers (Actors); (a) detailing the way of perceptual evaluation of actor's voice (speaking voice), (b) correlation of vocal projection with acoustic parameters and (c) effects of voice pathology on quality of life of the actors. Actors voice is not much researched upon in India, the professional voice researches in India are mostly concentrated on acoustical and perceptual analysis of singers. ,, Call center employees,  teachers. , Actor's voice is one of the cornerstone attributes of their career. Thus, the quality of their voice and changes in their voice may affect their career more so than many other occupations might. Research needs to be carried out to find out whether the actors are special as compared to non-actors, in the strategies they use to project their voices in Indian contexts. Hence the present study was aimed at analyzing acoustic and perceptual parameters for voice projection in actors and non-actors at three loudness levels (Habitual, moderate and stage performance level) and also to find out correlation between them.
| Subjects and Methods|| |
A total of 18 native Marathi actors which consists of 4 females and 14 males with a mean age of 32 years, with minimum 5 years of experience in theatre, formal training and no history of voice pathology participated in the study. Age, gender, language matched non-actors with a mean age of 31 years with no formal training, having no history of voice pathology were included in the study. Participants below 19 years and above 60 years were not included to rule out voice maturity and/or age related changes. Detailed history was taken from the participants (actors) of the study which include information about training of theater, voice care, vocal load, professional status, performance commitments, amount and nature of the voice training if any and performance environment etc., Summary of those case history details are given below in [Table 1].
Participants (actors and non-actors) were instructed to perform one dialogue at three different loudness level (Habitual, moderate and Stage performance level). The sample was recorded with the Digital recorder (Transcend with inbuilt microphone) was hanged in the participant's neck with the distance of 10 cm from mouth. Sample was then analyzed in Voice analysis software (Dr. Speech, Real Analysis-Tiger Electronics, Version 5) using LTAS.
The sample was analyzed and the groups were compared on the to study the voice projections on three different levels on the parameters: Average fundamental frequency (f0), mean sound pressure level (SPL), difference between mean frequency of strongest peak at 0-1 kHz and 3-4 kHz, level difference between the SPL of the strongest peak between 0-1 kHz and 3-4 kHz and pattern of spectral slope.
Samples were analyzed by two speech language pathologists perceptually on four parameters: projections, emotions, resonance and diction quality on a three point scale i.e,-1 for gentle projection, less or none emotions, hyponasal voice quality, poor diction 0.0 for adequate projection, expressive emotionality, clear diction and normal resonance quality and +1 for hard projection, over expressive emotionality, hypernasality, exaggerated diction. The scale is given in appendix II.
| Results|| |
Independent sample t-test was performed to see the difference between actors and non-factors on the acoustical parameters.
The actors had significantly higher average f0 than non-actors. The difference between the two groups was significant (P < 0.05) for habitual and moderate level, but strongly significant (P < 0.01) at stage performance level as shown in [Figure 1] and [Table 2], [Table 3], [Table 4].
A raising pattern was observed in mean SPL values for actors from habitual level to stage performance level. Though there was no significant difference between the mean SPL values between actors and non-actors at all the three projection levels as shown in [Figure 2] and [Table 5], [Table 6], [Table 7].
Difference between mean frequency of strongest peak at 0-1 kHz and 3-4 kHz [Figure 3].
|Figure 3: Difference between mean frequency of strongest peak at 0-1 kHz and 3-4 kHz|
Click here to view
There was statistical significant difference between the groups for moderate level and stage performance level. The actors had lower values than non-actors for all three levels as shown in [Table 8], [Table 9], [Table 10].
|Table 10: Statistical values for frequency difference at stage performance level|
Click here to view
Level difference between the SPL of the strongest peak between 0-1 kHz and 3-4 kHz: Though there was the difference between the means of both the groups' at all three projection levels, the difference was statistically significant only for moderate level (P < 0.05) and stage performance level (P < 0.05) as shown in [Table 11], [Table 12], [Table 13]. The average amplitude for actors for the strongest peak between 3 and 4 kHz was-24.7 dB at habitual level, -30.6 dB at a moderate level, -32.1 dB at stage performance level. And that of non-actors was -12.91 at a habitual level, -16.5 at a moderate level and 24.2 at stage performance [Figure 4].
|Table 13: Statistical values for level difference at stage performance level|
Click here to view
|Figure 4: Statistical diagram for level difference between the sound pressure level of the strongest peak between 0-1 kHz and 3-4 kHz|
Click here to view
Pattern of spectral slope: Pattern of spectral slope was found sharply falling with higher energy concentration in actors at 3-4 kHz and also 2-3 kHz unlike non-actors where the spectral slope was gradually falling and/or flat.
Client were rated on the given rating scale by two trained SLPs. Spearman's Inter-rater reliability quotient showed moderate correlation (r = 0.36). The actor's voice was rated better projected, with appropriate emotions and clearer diction than non-actors' voices. This indicates that there was consistency in rating in their evaluation. Perceptual parameters correlated moderately only with average f0 in stage performance level (r = 0.47).
| Discussion|| |
According to the results, it is clear that actors' voices are projected more efficiently than non-actors. Actors had higher f0 than their age and gender matched controls at all levels unlike what has been reported by other authors.  Raised f0 can be directly related to the higher first formant (Strongest peak between 0 and 1 KHz) which might be the result of narrowing of epilaryngeal tube as previously reported by authors.  This shows that actors may be using different supralaryngeal adjustments than non-actors. Mean SPL values for actors and non-actors were not significantly different at all three levels, which suggest that other factors than SPL contributed to the perception of loudness and projection. Mean level difference between SPL of the strongest peak for higher in actors than non-actors, which indicated there is no significant clustering of higher formants in actors. This probably is the result of raised f0 and higher mean SPL values in the region of 3-4 kHz in actors than non-actors. It is interesting to know that difference in mean frequency of strongest peak between 0-1 kHz and 3-4 kHz is significantly lower in actors than non-actors and the mean frequency between 3 and 4 kHz was significantly lower in actors than non-actors. This suggests that non-actors had faster glottal closing and stronger adduction than actors.  Hence perceptually non-actor's voices were also projected as harder than actors at forte level. Looking at the spectral slope of actors and non-actors, it is clear that effect of 2-3 kHz region was most important for the energy concentration between 3 and 4 kHz. Some of the participants had higher amplitude and stronger peak between 2 and 3 kHz than compared to 3 and 4 kHz especially in actors for all the three levels. This suggests that the higher amplitude in 3-4 kHz region is not the result of formant clustering, but more because of F2 and F3 of the actors. The result indicates that actors use supralaryngeal adjustments more efficiently than non-actors. There is also some chance of having better articulatory adjustments to project the voice more efficiently (perceptually not sounding hard). The non-actors had more of flat spectral slope and low energy levels both at 2-3 kHz and 3-4 kHz region, indicating lack of proper supralaryngeal adjustments.
| Conclusion|| |
Actors are efficient in projecting their voice perceptually than non-actors, which is in correlation with acoustic parameters at stage performance level, moderate level but the difference was not found at habitual levels. It appears that acoustically the projections are the result of favorable supralaryngeal adjustment in actors than non-actors. There is a need to further evaluate the effect of years of experience, gender, age and training on vocal projection in actors.
| Acknowledgments|| |
We sincerely thank the actors and non-actors for contributing to this paper as participants. We also thank Dr. Gauri Belsare, Principal and Dr. C.S. Vanaja, HOD, School of Audiology and Speech Language Pathology, Bharati Vidyapeeth Deemed University, Pune for giving the permission to conduct the study in the department. We also thank faculty members of School of Audiology and Speech Language pathology for helping us in carrying out perceptual evaluation of the samples.
| References|| |
|1.||Koufman JA, Isaacson G. The spectrum of vocal dysfunction. In: Koufman J, Isaacson G, editors. The Otolaryngologic Clinics of North America: Voice Disorders. Philadelphia: W.B. Saunders; 1991. p. 47. |
|2.||Shirvastav R, Wingate J. Perceptual voice and assessment of singers' voice. In: Benninger MS, Murry T, editors. The Singer's Voice. San Diego: Plural Publishing Inc.; 2007. p. 64-77. |
|3.||Hirano M. Clinical Examination of Voice. NY: Springer Wien; 1981. |
|4.||Zraick RI, Kempster GB, Connor NP, Thibeault S, Klaben BK, Bursac Z, et al. Establishing validity of the consensus auditory-perceptual evaluation of voice (CAPE-V). Am J Speech Lang Pathol 2011;20:14-22. |
|5.||Master S, De Biase N, Chiari BM, Laukkanen AM. Acoustic and perceptual analyses of Brazilian male actors' and nonactors' voices: Long-term average spectrum and the "actor's formant". J Voice 2008;22:146-54. |
|6.||Leino T. Long-term average spectrum study on speaking voice quality in male actors. In: Friberg A, Iwarsson J, Janson E, Sundberg J, editors. SMAC03 (Proceedings of the Stockholm Music Acoustics Conference). Stockholm: Royal Swadish Academy of Music; 1994. p. 206-10. |
|7.||Pinczower R, Oates J. Vocal projection in actors: The long-term average spectral features that distinguish comfortable acting voice from voicing with maximal projection in male actors. J Voice 2005;19:440-53. |
|8.||Sundberg J. The singer's formant revisited. Voice 1995;4:106-19. |
|9.||Nawka T, Anders LC, Cebulla M, Zurakowski D. The speaker's formant in male voices. J Voice 2005;19:440-53. |
|10.||Bhoominathan P. Characteristics required singing-observations from south Indian classical singers. Proceedings of National Seminar on Professional Voice Care, AIISH - RCI-CRE. Mysore, India: 2005. |
|11.||Bhoominathan P. Do Carnatic singers have singer's formant? An exploratory study. Proceedings of the Frontiers of Research on Speech and Music (FRSM). Chidambaram, India: 2004. p. 22-6. |
|12.||Supraja A, Savithri SR. Relationship between habitual frequency of voice and 'a: dha: rasruti' in Carnatic vocal music. Proceeding of the Frontiers of Research on Speech and Music (FRSM). Mysore, India; 2007. p. 145-7. |
|13.||Shah, Sanghi. A study of voice characteristics and reflux symptom index in professional voice users (call centre employees). J Indian Speech Hear Assoc 2010;24:39-47. |
|14.||Behra S, Savithri SR. Voice characteristics of prospective and professional teachers. JISHA 2005;19:62-7. |
|15.||Raja SR, Savithri SR. Voicing periods in primary school teachers. JAIISH 2009;28:36-41 |
|16.||Gauffin J, Sundberg J. Spectral correlates of glottal voice source waveform characteristics. J Speech Hear Res 1989;32:556-65. |
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7], [Table 8], [Table 9], [Table 10], [Table 11], [Table 12], [Table 13]