Abstract: 10 young female singing students were compared to their teachers in a vocal pitch-matching task with three standardized intervals (third, a fifth and an octave) and in two conditions (with / without piano modelling). Period durations at the offset of the base tone and at the onset of the target tone were analyzed in detail. Contrary to untrained individuals, the singing students appear to have already reached a degree of neuromuscular control in pitch matching comparable to that of professionals. There is also no effect of the interval extent on the accuracy of pitch-matching, and the reference of the piano does not play a significant role. In general, a Fo-instability characterizes the first cycles of the target tone; it could be explained by mechanical readjustments. However, the periodicity of the last cycles of the base tone is also significantly perturbed at the moment at which the subject anticipates the pitch jump, and this perturbation increases with the extent of the pitch jump to be achieved. This perturbation plausibly reflects an equivalent of the pre-phonatory burst of asynchronous muscle action potentials observed before the onset of phonation. However the phenomenon is so fast that it is not perceived by the listener.
Keywords: pitch- matching, singing, perturbation, anticipation, M Vocalis.
I.INTRODUCTION
Trained singers perform more accurately than untrained individuals in a pitch-matching task, consisting of matching the vocal fundamental frequency to a preset target tone [1-2]. This ability seems specific to
neuromuscular voice regulation. Murry and Caligiuri [3] have compared the performances of adult singers and adult non-singers on an auditory-motor task (pitch matching) and a visuomotor task (turning the wrist to match a visual target). Singers performed more accurately in the pitch-matching task than non-singers, and less experienced singers were less consistent during pitch-matching tasks than more experienced ones, but wrist targets were achieved with comparable efficiency. These results suggest that specific neuro-muscular training and specific experience with motor planning play a key role in technical aspects of singing abilities; however, the importance of their influence is not known. Does a singing student already master this ability quite early, or is it only acquired after years of practice at a professional level? Other relevant aspects that pertain to the accuracy of the neuromuscular control are the influence of modelling the target tone by providing the note with a piano (auditory information), and the effect of the extent of the interval. To address these questions, young conservatory students were compared with their teachers in a pitch matching experiment with three different intervals and two conditions: with or without modelling.
To get the best possible insight in the control mechanism of pitch adjustment, all sound periods in the critical phases (first cycles of target tone and last cycles of base tone) were measured period by period. Voice onset is known to show typically much greater frequency perturbation than the steady state midportion of a sustained vowel [4]. An additional aim of the study was to compare this short-term instability in our two groups of singers.
II. MATERIAL & METHODS
ANTICIPATION OF A NEUROMUSCULAR TUNING IN M. VOCALIS
PERTURBS THE PERIODICITY OF VOCAL FOLD VIBRATION: THE
UNEXPEXTED FINDING OF A PITCH-MATCHING EXPERIMENT
COMPARING SINGING STUDENTS WITH HIGH-LEVEL
PROFESSIONALS
P.H. Dejonckere
1, J. Lebacq
2, C. Manfredi
31 University of Leuven, Neurosciences, Exp. ORL, Belgium ; Federal Institute of Occupational Diseases, Av. de l’Astronomie, 1, B-1210 Brussels Belgium.
2 Institute of Neuroscience CEMO, Université Catholique de Louvain, Avenue Hippocrate 55 bte B1.55.12B-1200 Brussels, Belgium.
3 Department of Information Engineering, Università degli Studi di Firenze, Via S. Marta 3, 50139 Firenze, Italy.
Claudia Manfredi (edited by), Models and analysis of vocal emissions for biomedical applications : 8 th international workshop : December 16-18, 2013 ISBN 978-88-6655-469-1 (print) ISBN 978-88-6655-470-7 (online) © 2013 Firenze University Press
0 0,5 1 1,5 2 2,5 3 3,5
Base Target start Target later Third Fifth Octave
Students
Coefficient of variation(%)
Fig. 3: Fo perturbation in the last 10 cycles of the base tone, in the first 5 cycles of the target tone and in the 10 subsequent cycles of the target tone. (Singing students)
Fig. 4: Fo perturbation in the last 10 cycles of the base tone, in the first 5 cycles of the target tone, and in the 10 subsequent cycles of the target tone. (Professional singers)
IV. DISCUSSION
The hypothesis that professional classical singers achieve a more precise neuromuscular control in pitch tuning than young students is to be rejected. This suggests that this ability does not require a very large amount of training or experience.
In the categories of subjects investigated in this work, modelling of tone intervals by a piano seems ineffective. As expected, the frequency perturbation is, for all conditions, much larger during the first cycles of the
target tone than in the ten subsequent cycles. This phenomenon is plausibly related to a necessary mechanical readjustment and stabilization of muscle tension and to a new balance between glottal resistance and lung pressure. Here again no difference is found, neither between groups nor between intervals.
Most intriguing is the period perturbation induced by the anticipation of the pitch jump. The jitter phenomenon is still incompletely understood, but it is usually explained by mechanical factors within the vocal fold tissues (viscosity, non-linearity), by internal body noise (pulsations in blood vessels) and by asynchronous firing of motor units in the M. vocalis and cricothyroideus. The third factor is probably mainly directly involved in the increased Fo perturbations observed in the present experiment.
The pre-phonatory tuning of the M. vocalis is a well known phenomenon and can be clearly observed using electromyography: the thyroarytenoid muscles show a burst of muscle action potentials preceding the onset of phonation followed by a decreased activity [6]. The time interval between the start of the change in electrical activity and the onset of voice emission is 100 to 200 ms (actually from 50 to 600 ms) [7]. This enhanced pre-phonatory activity consists of firing of motor units within the M. vocalis and these action potentials are asynchronous with the vocal fold vibration frequency. This could explain the perturbation effect on the regularity of this vibration frequency. As there is a clear positive correlation in modal register between voice pitch and muscle tension in the M. vocalis [8], it may be expected that the importance of such a pre-target tuning activity is dependent on the extent of the ascending tone interval. Consequently, it may be hypothesized that it interferes accordingly with the tension regulation during the last ms of the base tone.
V. CONCLUSION
Young singing students appear to have already reached a degree of neuromuscular control in pitch matching comparable to that of true professionals.
For a mild range of tone intervals (third – fifth – octave) there is no effect on the accuracy of pitch-matching. Reference of the piano does not play a significant role either.
In general, a Fo-instability characterizes the first cycles of the target tone and it could be explained by the mechanical readjustments. However the last cycles of the base tone are also perturbed at the moment the subject anticipates the pitch jump and the perturbation increases with the extent of the pitch jump that will be achieved. This perturbation plausibly reflects an equivalent of the pre-phonatory burst of asynchronous muscle action potentials observed before the onset of phonation. Subjects were 10 healthy female singing students (ages
18-22) on the one hand, and 10 healthy female professional classic singers (ages 26-43) on the other hand, all of them being also singing teachers in three different conservatories. They were asked to produce 10 times 3 standard intervals (third, fifth and octave) starting from a same base tone d1 (~ 294 Hz) that was given by a piano. In the five first trials, the interval was modelled by the piano, while in the last five ones only the start tone was provided. The subjects were requested to sing on /a/ at comfortable loudness level without vibrato and with a short interruption between the two tones of each interval (thus not legato and without portamento). All recordings were made digitally in a quiet room (conservatory classroom), with a Sennheiser MD 421 N microphone at 30 cm from the mouth and at a sample frequency of 44.100 Hz. The PRAAT program (PRAAT 5.3.10, 2012 by P. Boersma & D. Weenink: www.praat.org.) was used for displaying and stretching the oscillograms. Period duration measurements were made cycle by cycle with cursors on the computer screen on the last ten cycles before the interval, and on the first fifteen cycles after the pitch jump (Fig. 1). This procedure is very easy in normal voice signals, and has been checked in another context [5].
Principle of measurement
time Fo 10 cycles 5 10 cycles Tone intervalFig. 1: Pitch-matching task for a given interval. Principle of measurement. X- axis is time, Y-axis is fundamental frequency. At a given moment the singer makes a pitch jump (third / fifth or octave) without legato, thus with a short interruption in voice emission. Cycle duration is measured just before (10 cycles) and just after (5 & 10 cycles) the jump.
III. RESULTS
Fig. 2 shows the Fo quotients for the three intervals (in average 1.2, 1.5 and 2), with their standard deviations, in
both professionals and in students, with or without modelling of the interval at the piano.
The standard deviations (SD) of the quotients reflect the accuracy of the interval. A null SD should indicate that the singer always realizes exactly a frequency quotient of 1.2, 1.5 or 2. No systematic deviations were observed in the sense of making the interval smaller or larger. The average standard deviations of the quotients do not significantly differ for the three intervals.
No significant differences were observed between students and teachers, and there is no significant effect of modelling by the piano.
±Std. Dev. ±Std. Err. Mean Quotient of frequencies : target / base
pr of . 1.1 1.4 1.7 2 2.3 with model st ud . 1.1 1.4 1.7 2 2.3
third fifth octave
without model third fifth octave
Fig. 2: Fo quotients for the three intervals (in average 1.2, 1.5 and 2) with their standard deviations in professionals and in students, with and without modelling of the interval at the piano.
In both groups of subjects, for all three intervals and for the two modelling conditions, there is a significant increase in Fo-perturbation (variation coefficient of period) in the first five cycles of the target tone when compared with the 10 subsequent cycles (always p < .001). However there is no difference between singing students and singing teachers, and no significant effect of the extent of the interval (Figs. 3 & 4).
Interestingly, the last ten cycles before the interval also demonstrate a significant increase in Fo-perturbation compared to the steady state phonation (p<0.001). This is observed in both groups of subjects. Furthermore, the variation coefficient of these last ten cycles significantly increases (p < 0.01) with the extent of the interval (Figs. 3 & 4).
0 0,5 1 1,5 2 2,5 3 3,5
Base Target start Target later Third Fifth Octave
Students
Coefficient of variation(%)
Fig. 3: Fo perturbation in the last 10 cycles of the base tone, in the first 5 cycles of the target tone and in the 10 subsequent cycles of the target tone. (Singing students)
Fig. 4: Fo perturbation in the last 10 cycles of the base tone, in the first 5 cycles of the target tone, and in the 10 subsequent cycles of the target tone. (Professional singers)
IV. DISCUSSION
The hypothesis that professional classical singers achieve a more precise neuromuscular control in pitch tuning than young students is to be rejected. This suggests that this ability does not require a very large amount of training or experience.
In the categories of subjects investigated in this work, modelling of tone intervals by a piano seems ineffective. As expected, the frequency perturbation is, for all conditions, much larger during the first cycles of the
target tone than in the ten subsequent cycles. This phenomenon is plausibly related to a necessary mechanical readjustment and stabilization of muscle tension and to a new balance between glottal resistance and lung pressure. Here again no difference is found, neither between groups nor between intervals.
Most intriguing is the period perturbation induced by the anticipation of the pitch jump. The jitter phenomenon is still incompletely understood, but it is usually explained by mechanical factors within the vocal fold tissues (viscosity, non-linearity), by internal body noise (pulsations in blood vessels) and by asynchronous firing of motor units in the M. vocalis and cricothyroideus. The third factor is probably mainly directly involved in the increased Fo perturbations observed in the present experiment.
The pre-phonatory tuning of the M. vocalis is a well known phenomenon and can be clearly observed using electromyography: the thyroarytenoid muscles show a burst of muscle action potentials preceding the onset of phonation followed by a decreased activity [6]. The time interval between the start of the change in electrical activity and the onset of voice emission is 100 to 200 ms (actually from 50 to 600 ms) [7]. This enhanced pre-phonatory activity consists of firing of motor units within the M. vocalis and these action potentials are asynchronous with the vocal fold vibration frequency. This could explain the perturbation effect on the regularity of this vibration frequency. As there is a clear positive correlation in modal register between voice pitch and muscle tension in the M. vocalis [8], it may be expected that the importance of such a pre-target tuning activity is dependent on the extent of the ascending tone interval. Consequently, it may be hypothesized that it interferes accordingly with the tension regulation during the last ms of the base tone.
V. CONCLUSION
Young singing students appear to have already reached a degree of neuromuscular control in pitch matching comparable to that of true professionals.
For a mild range of tone intervals (third – fifth – octave) there is no effect on the accuracy of pitch-matching. Reference of the piano does not play a significant role either.
In general, a Fo-instability characterizes the first cycles of the target tone and it could be explained by the mechanical readjustments. However the last cycles of the base tone are also perturbed at the moment the subject anticipates the pitch jump and the perturbation increases with the extent of the pitch jump that will be achieved. This perturbation plausibly reflects an equivalent of the pre-phonatory burst of asynchronous muscle action potentials observed before the onset of phonation.
Abstract: Singers constitute a specific population sensitive to vocal disability, which may have a higher impact on their quality of life compared to non-singers. A specific questionnaire, the Singing Voice Handicap Index (SVHI) was created and validated aimed to measure the physical, social, emotional and economic impacts of voice problems on the lives of singers. Aim of this study was to validate the Italian version of the SVHI. The validated English version of the SVHI was translated into Italian and then discussed with several voice care professionals. The Italian version of the SVHI was administered to 214 consecutive singers (91 males and 123 females, mean age: 32.62±10.85). Voice problem complaints were expressed by 97 of the singers, while 117 were healthy and had no voice conditions. All subjects underwent a phoniatric consultation with videolaryngostroboscopy to ascertain the condition of the vocal folds. Internal consistency of the Italian version of the SVHI showed a Cronbach’s α of 0.97. The test-retest reliability was assessed by comparing the responses obtained by all subjects in two different administrations of the questionnaire; the difference was not significant (p=ns). The SVHI scores in healthy singers was significantly lower than the one obtained in the group of singers with a vocal fold abnormality (29.26±25.72 and 45.62±27.95, p<0.001, respectively). The Italian version of the SVHI was successfully validated as a suitable instrument for the self-evaluation of handicaps related to voice problems in the context of singing.
Keywords : Singing, Voice Disturbance, Questionnaire Design, Self Report
I.INTRODUCTION
Self-administered questionnaires are used to assess the impacts of health problems on the quality of life of patients: a disability in performing a daily task, defined as a handicap, could cause a disadvantage in social, economic or environmental aspects of life [1]. Several questionnaires have been designed to measure the impact of voice problems on the lives of individuals: the most
popular is the Voice Handicap Index (VHI) [2], which has been validated and translated into several languages [3-5]. The VHI was developed to assess the subjective perception of disability related to voice disorders in all types of patients [2].
Singers constitute a specific population of professionals particularly at risk for voice problems. The perception of a voice problem in singing is often related to specific symptoms, such as difficulty in the passaggio, vocal endurance and diminished range [6], aspects that are not assessed by the VHI. Furthermore, singers are often more sensitive to vocal disabilities, which may have a higher impact on their quality of life compared to non-singers [7]. Hence, to obtain a self-assessing instrument able to evaluate vocal disability in singers, in 2007, Cohen et al. created and validated a specific questionnaire, the Singing Voice Handicap Index (SVHI), aimed to measure the physical, social, emotional and economic impacts of voice problems on the lives of singers [8]. The SVHI is a 36-item self-administered questionnaire that is able to assess difficulties related to voice health status typical of the singing professional. The items address symptoms frequently reported to laryngologists and speech pathologists by singers. Among singers, the SVHI is also more sensitive to clinical changes than the VHI [9], which proves the validity of the SVHI in measuring treatment outcomes in the singing population. The aim of this study was to validate the Italian version of the VSHI.
II. METHODS
Development of the Italian version of the SVHI: An
Italian translation of the validated English version of the SVHI was carried out by a qualified professional translator. The first version of the questionnaire was then discussed by 2 phoniatricians, 2 speech therapists, 2 singing teachers and 2 professional singers to improve the translation and to make it more understandable to singers. After, the new Italian version was to re-translated in English and, finally, re-translated in Italian language. Each of the 36 items of the questionnaire was
VALIDATION OF THE ITALIAN VERSION OF THE SINGING VOICE
HANDICAP INDEX
G. Baracca
1, G.Cantarella
1, S. Forti
2, F. Fussi
31 Otolaryngology Department, Fondazione IRCCS Cà Granda Ospedale Policlinico, Milan, Italy,
giovanna.baracca@gmail.com
2 Audiology Unit,Fondazione IRCCS Cà Granda Ospedale Policlinico, Milan, Italy aut_est@yahoo.it 3AUSL di Ravenna, Ravenna, Italy ffussi@libero.it
However the phenomenon is so fast that it not perceived by the listener.
REFERENCES
[1] T. Murry,” Pitch-matching accuracy in singers and non-singers.” J Voice, vol. 4, pp. 317-321, 1990. [2] J.M. Estis, A. Dean-Claytor, R.E. Moore and T.L.
Rowell, “Pitch-matching accuracy in trained singers and untrained individuals: the impact of musical interference and noise. “ JVoice vol. 25, pp. 173-180, 2011.
[3] T. Murry, MP Caliguiri, “Phonatory and nonphonatory motor control in singers.” J Voice, vol. 3, pp. 257-263, 1989.
[4] Y. Koike, “Application of some acoustic measures for the evaluation of laryngeal dysfunction”, Studia Phonologica, vol. 7, pp. 17-23, 1973. [5] P.H. Dejonckere, A. Giordano, J. Schoentgen, J.
Fraj, L. Bocchi, C Manfredi,, “To what degree of voice perturbation are jitter measurements valid? A novel approach with synthesized vowels and visuo-perceptual pattern recognition”, Biomedical Signal Processing and Control, Vol. 7, pp.: 37-42, 2012.
[6] A.D. Hillel, “The study of laryngeal muscle activity in normal human subjects and in patients with laryngeal dystonia using multiple fine-wire electromyography”, Laryngoscope, Vol. 111, 2 Suppl., pp. 1- 47, 2001.
[7] P.H. Dejonckere, EMG of the larynx. Marc Pietteur: Liège ISBN 2-87211-000-3, 1987. [8] P.H. Dejonckere, “Les mécanismes musculaires
élémentaires de régulation de la tension de la corde vocale au cours de la phonation”, Folia Phon. vol. 32, pp. 1-13, 1980.
