ArticlePDF AvailableLiterature Review

Abstract and Figures

This study, focused on the laryngeal source level, introduces the concept of laryngeal vibratory mechanism. Human phonation is characterized by the use of four laryngeal mechanisms, labeled M0-M3, as evidenced by the electroglottographic (EGG) study of the transition phenomena between mechanisms with a population of men and women, trained and untrained singers. Macroscopic and local descriptions of the EGG signal are analyzed during the production of glissandos and held notes with different mechanisms. The transition from one mechanism to another of higher rank is characterized by a jump in frequency, a reduction of EGG amplitude, and a change in the shape of the derivative of the EGG (which may correspond to a reduction of the vibratory mass). These characteristics are used to identify a transition between two mechanisms, in complement with acoustic spectrographic analyses. The pitches of transitions between the two main mechanisms M1 and M2 and the range of the frequency-overlap region are described in detail. The notion of vocal register is revisited in the light of these concepts of laryngeal mechanism. The literature on vocal registers is reviewed, and it is shown that the confusion often cited with respect to this notion may be related to the heterogeneity of the approaches and methods used to describe the phenomena and to the multiplicity of descriptors. Therefore, the terminology of the registers is organized depending on their relation to the four laryngeal vibratory mechanisms.
Content may be subject to copyright.
A preview of the PDF is not available
... cycles. This case undeniably has to be considered as creak (phonation mechanism zero in terms of the classification of Roubeau, Henrich, and Castellengo (2009)). Overall, this example can be described as an extreme sample of clear lapse into creaky voice, with the lowest possible f 0 dEGG and O q dEGG , but still with a single pulse per cycle. ...
... The most obvious feature of double-pulsed creak is the alternation of short and long cycles (the starting cycle can be either a % short or a long cycle) in a great change of fundamental frequency. This characteristic has been broadly recognized by numerous previous works such as: "double pulsation" in (Paul Moore and Leden, 1958), "double glottic pulses" in (Hollien and Ronald W. Wendahl, 1968) and (Hedelin and Huber, 1990), (Roubeau, Henrich, and Castellengo, 2009), etc. The other also mentioned the possibilities of triplet cycles (Blomgren et al., 1998) or higher multiples (Keating, Garellek, and Kreiman, 2015). ...
... The retained values are lower than 30% and can reach extremely low values: below 15% in some speakers. These values offer telltale evidence of the presence of creaky voice (phonation mechanism zero according to the classification ofRoubeau, Henrich, and Castellengo 2009). ...
Thesis
All languages in the Vietic subbranch of Austroasiatic have at least one glottalized tone. This thesis zooms in on one of these languages: Muong (in Vietnamese orthography: Mường, endonym: /mon³/), spoken in Kim Thuong (Phu Tho, Vietnam). Twenty speakers recorded twelve tonal minimal sets of the five tones of smooth syllables, plus three tonal minimal pairs of the two tones of checked syllables, under two conditions: in isolation and in a carrier sentence. Acoustic and electroglottographic recordings allow for estimating fundamental frequency, glottal open quotient and duration. These parameters are compared across tones, experimental conditions and speakers, in order to contribute to a better understanding of glottalization as a feature of linguistic tones. First, allotones of the phonologically glottalized tone in Muong (Tone 4) are classified on a phonetic basis, confirming the consistent presence of creak. It is tempting to contrast it with the glottally constricted tones of Northern Vietnamese (with which Muong is in sustained language contact). However, the phonological discussion emphasizes that analysis of Tone 4 as a prototypical "creaky tone" would be a pitfall. Tone 4 behaves in key respects like the other tones in the system: it is not defined solely by phonation type. Moreover, the range of phonetic (allotonic) variation of Tone 4 includes cases of glottal constriction. Use of a phonetic nomenclature for types of glottalization serves as a basis for describing the interaction of glottalization with intonation.
... The phonatory stability of vocal fold oscillation has mostly been investigated in regard of voluntary changes in vibratory mode by adjusting muscles, as they occur for instance during professional singing when switching between registers [9][10][11][12]. However, there are also spontaneous changes in vocal fold vibratory mode that can lead to frequency drops or jumps. ...
... That, in the mode prior to each transition, the following is observed: (a) a relatively high fundamental frequency (usually around 400 Hz), (b) relatively weak harmonics, and (c) a roughly equal OD and CD. These are characteristics generally associated with the falsetto register in the human singing voice [9][10][11][12]. On the other hand, in the mode subsequent to each transition, the following is observed: (a) a relatively low fundamental frequency (usually around 200 Hz,), (b) relatively strong harmonics, (c) a CD which is much longer than the OD. ...
... On the other hand, in the mode subsequent to each transition, the following is observed: (a) a relatively low fundamental frequency (usually around 200 Hz,), (b) relatively strong harmonics, (c) a CD which is much longer than the OD. These are characteristics generally associated with the chest register in the human singing voice [10][11][12]. Thus, the spontaneous jumps we have observed in the study appear to be consistent with the falsetto-to-chest register transition frequently observed in the human singing voice. ...
Article
Full-text available
Investigations of neuromuscular control of voice production have primarily focused on the roles of muscle activation levels, posture, and stiffness at phonation onset. However, little work has been done investigating the stability of the phonation process in regards to spontaneous changes in vibratory mode of vocal fold oscillation as a function of neuromuscular activation. We evaluated 320 phonatory conditions representing combinations of superior and recurrent laryngeal nerve (SLN and RLN) activations in an in vivo canine model of phonation. At each combination of neuromuscular input, airflow was increased linearly to reach phonation onset and beyond from 300 to 1400 mL/s. High-speed video and acoustic data were recorded during phonation, and spectrograms and glottal-area-based parameters were calculated. Vibratory mode changes were detected based on sudden increases or drops of local fundamental frequency. Mode changes occurred only when SLNs were concurrently stimulated and were more frequent for higher, less asymmetric RLN stimulation. A slight increase in amplitude and cycle length perturbation usually preceded the changes in the vibratory mode. However, no inherent differences between signals with mode changes and signals without were found.
... The loudness is mostly controlled by the lung pressure and the glottal resistance. The different registers are believed to originate from the different vibratory patterns of the vocal folds (also referred to laryngeal mechanisms, [Henrich, 2006;Roubeau et al., 2009]. This is evidenced by experiments with excised larynges where voice of two registers can be produced without the vocal tract (e.g., [Alipour and Scherer, 2007;Baer, 1969;Van den Berg, 1963;Müller and Baly, 1848]). ...
... [ Roubeau et al., 2009] introduced the concept of laryngeal vibratory mechanism based on electroglottographic study of the transition phenomena between registers. The transition from one mechanism to another of higher rank is characterized by a jump in frequency, a reduction of EGG amplitude, and a change in the shape of the derivative of the EGG, indicating a change of vibratory mechanism. ...
Article
While many may take it for granted, the human voice is an incredible feat. An average person can produce a great variety of voices and change voice characteristics agilely even without formal training. Last several decades of research has established that the production of voice is largely a mechanical process: i.e., the sustained vibration of the vocal folds driven by the glottal air flow. Since one only has a single pair of vocal folds, the versatility comes with the ability to change the mechanical status of the vocal folds, including vocal fold length and thickness, tension, and level of adduction, through activation of the laryngeal muscles. However, the relationship between laryngeal muscle activity and the characteristics of voice is not well understood due to limitations in experimental observation and simplifications in modelling and simulations. The science is still far behind the art. The current research aims to investigate first the relationship between laryngeal muscle activation and the posture of the vocal folds and second the relationship between voice source characteristics and vocal fold mechanical status using more comprehensive numerical models and simulations, thus improving the understanding of the roles of each laryngeal muscle in voice control. To do so, (1) the mechanics involved in vocal fold posturing and vibration, especially muscle contraction; (2) the realistic anatomical structure of the larynx must be considered properly. To achieve this goal, a numerical model of the larynx as realistic as possible was built. The geometry of the laryngeal components was reconstructed from high resolution MRI (Magnetic Resonance Imaging) data of an excised canine larynx, which makes more accurate the representation of the muscles and their sub-compartments, cartilages, and other important anatomical features of the larynx. A previously proposed muscle activation model was implemented in a 3D finite element package and applied to the larynx model to simulate the action of laryngeal muscles. After validation of the numerical model against experimental data, extensive parametric studies involving different combination of muscle activations were conducted to investigate how the voice source is controlled with laryngeal muscles. In the course of this study, some work was done to couple the same finite element tool with a Genetic Algorithm program to inversely determine model parameters in biomechanical models. The method was applied in a collaborated study on shape changes of a fish fin during swimming. This study is presented as a separate chapter at the end of this thesis. The method has potential application in determining parameters in vocal fold models and optimizing clinical vocal fold procedures. This thesis is essentially an assembly of the papers published by the author during the doctoral study, with the addition of an introductory chapter. Chapter 1 reviews the overall principles of voice production, the biomechanical basis of voice control, and past studies on voice control with a focus on the fundamental frequency. Chapter 2 describes the major numerical methods employed in this research with an emphasis on the finite element method. The muscle activation model is also described in this chapter. Chapter 3 describes the building of the larynx model from MRI data and its partial validation. Chapter 4 presents the application of the larynx model to posturing studies, including parametric activation of muscle groups and specific topics related to vocal fold posturing. Chapter 5 describes the change of vocal fold vibration dynamics under the influence of the interaction of the cricothyroid muscle and the thyroarytenoid muscle. The Flow-structure interaction simulations was realized by coupling the larynx model to a simple Bernoulli flow model and a two-stage simulation technique. Chapter 6 concludes the current thesis study. Suggestions for future studies are proposed. Chapter 7 is an independent study that is not related to voice control. It describes a numerical framework that inversely determines and validates model parameters of biomechanical models. The application of the proposed framework to a finite element model of a fish fin is presented.
... Arousal was also shown to bring most of the acoustic changes in a voice in [19] study. S3 appears to mostly rely on M1 vibration mechanism (modal), while S6 uses a M2 voice register (falsetto), with higher pitch (on registers: [39], [40]). ...
... Despite those characteristics, these two registers are assumed to be associated mostly with different laryngeal adjustments, which produce different vocal fold vibratory patterns. 4 Such laryngeal adjustments are sometimes referred to as "laryngeal mechanisms" and numbered from low pitch to high pitch (M0, M1, M2 and M3), 5 where M1 and M2 are usually associated with the chest and the head registers, respectively. In order to observe the activity of the laryngeal muscles when using different registers, Hirano et al, 6 performed electromyographic (EMG) measurements. ...
Article
Full-text available
Voice registers are assumed to be related to different laryngeal adjustments, but objective evidence has been insufficient. While chest register is usually associated with the lower pitch range, and head register with the higher pitch range, here we investigated a professional singer who claimed an ability to produce both these registers at every pitch, throughout her entire singing range. The singer performed separated phonations alternating between the two registers (further called chest-like and head-like) at all pitches from C3 (131 Hz) to C6 (1047 Hz). We monitored the vocal fold vibrations using high-speed video endoscopy and electroglottography. The microphone sound was recorded and used for blind listening tests performed by the three authors (insiders) and by six “naive” participants (outsiders). The outsiders correctly identified the registers in 64% of the cases, and the insiders in 89% of the cases. Objective analysis revealed larger closed quotient and vertical phase differences for the chest-like register within the lower range below G4 (<392 Hz), and also a larger closed quotient at the membranous glottis within the higher range above Bb4 (>466 Hz), but not between Ab4-A4 (415-440 Hz). The normalized amplitude quotient was consistently lower in the chest-like register throughout the entire range. The results indicate that that the singer employed subtle laryngeal control mechanisms for the chest-like and head-like phonations on top of the traditionally recognized low-pitched chest and high-pitched head register phenomena. Across all pitches, the chest-like register was produced with more rapid glottal closure that was usually, but not necessarily, accompanied also by stronger adduction of membranous glottis. These register changes were not always easily perceivable by listeners, however.
Thesis
Die vorliegende Studie diente der Erprobung der neuartigen Methode der quasi-dynamischen 3D-Magnetresonanztomographie und der Erweiterung der Erkenntnisse bezüglich Resonanzstrategien im Bereich des Pfeifregisters. Hinsichtlich der Resonanzstrategien verhielten sich die Probandinnen nicht einheitlich. Es lässt sich festhalten, dass im Bereich des Pfeifregisters keine eindeutige registerbezogene Vokaltraktanpassung mit Änderung der Tuning-Strategie stattfindet. Eine konstante Einstellung kann zu Kreuzungspunkten von Resonanzen und Teiltönen führen und damit verbunden auch zu einem Ende jeglichen Tunings. Das fR1:fo Tuning erwies sich als bis F6 (1397 Hz) anwendbar. Die quasi-dynamische 3D-Methode hat Vorteile gegenüber anderen Messweisen, insbesondere bzgl. des Zusammenhangs der Vokaltrakt-Einstellungen innerhalb einer sängerischen Phrase und der Einsparung von aufwendigem Druckmaterial und akustischen Messaufbauten. Im Detail zeigte sich noch Verbesserungsbedarf in Hinblick auf den Zeitaufwand für die Vokaltrakt-Segmentierung, welche weiter automatisiert werden sollte. Die Ergebnisse innerhalb der Methode erwiesen sich als konsistent, jedoch bestehen weiterhin Abweichungen ungeklärten Ursprungs im Vergleich zu anderen Methoden. Diesbezüglich sollten der Applikation VTTF zur Errechnung der Vokaltraktresonanzen weitere Variablen implementiert werden sowie weitere Methodenvergleiche durchgeführt werden. Es ergaben sich zahlreiche Ansätze für weiterführende Studien zu den Themen Nasalanz, subglottische Strukturen, Nicht-lineare Interaktionen, Formantbandbreiten in hohen Lagen, Bedeutung der fR2/2fo Kreuzung für den Pfeifregisterklang, sowie spektrale Ereignisse in Verbindung mit der Tonhöhe A5, dem Teiltonbereich 3 - 4,5 kHz, einem möglichen Rauschen und Subharmonischen.
Article
Riassunto Le pieghe vocali (nome anatomico ufficiale delle corde vocali) entrano in vibrazione nella laringe al passaggio dell’aria espirata. Questa vibrazione a sua volta provoca la vibrazione dell’aria contenuta nel tratto vocale, il che corrisponde alla comparsa del suono della voce. Esiste una parte attiva del fenomeno che è l’accostamento delle pieghe grazie alla contrazione dei muscoli della laringe il cui nervo motore è il nervo laringeo inferiore. Poi la vibrazione stessa è un fenomeno meccanico legato al passaggio dell’aria espirata attraverso le pieghe vocali accostate l’una contro l’altra. La struttura particolare della piega vocale e che rende possibile questa vibrazione è detta “lamina propria”.
Article
Aim This report aims to perform a preliminary evaluation of the role of taping in the improvement of phonasthenia among professional voice users. The larynx is a vital organ but also a work instrument for many people. Prevention of erroneous use of the vocal mechanism is essential for an optimal voice performance. Methods Nineteen singers complaining voice fatigue underwent taping application in peri-laryngeal areas of the anterior neck for 10 days. Taping effects were estimated by the evaluation of voice analysis parameters (jitter, shimmer, noise to harmonic ratio and singing power ratio) performed in M1 and M2 mechanisms with PRAAT software before and after treatment. In addition, a subjective assessment of phonasthenia was performed using a self-administrated questionnaire. Results Statistically significant reduction of Jitter% in M1 (P = 0.021) was reported at the end of treatment; also, an improvement of NHR in M2 was found immediately after the application (P = 0.012) and after 10 days (P = 0.002). These outcomes suggest possible beneficial effects of laryngeal taping on relieving voice fatigue, and therefore improving voice quality. Self-assessment results are consistent with this finding. Conclusion Since the application of muscular taping in phoniatrics is still in its early stage, further studies and in particular a standardized protocol of application, could facilitate the spread of this technique, also offering help in comparing results.
Article
Full-text available
Abstract From a physiological point of view, two main laryngeal vibratory mechanisms, M1 and M2, are used successively from the bottom to the top of the vocal range. From a musical point of view, singers distinguish many registers, most of which rely on resonance adjustements. Voix mixte, which is related to the area of overlap of M1 and M2, is a register found in different voice categories. Use of the voix mixte allows the singers to realize a homogeneous voice timber throughout their tessitura. Contradictory views have been expressed on the true nature of this register: laryngeal, or resonantial? The present study on French voix mixte, carried out with 5 professional singers of both sexes, shows, on the basis of glottal open quotient (Oq) measurements, that voix mixte is not related to a different, or "mixed", laryngeal mechanism. Voix mixte sounds are always clearly produced in a given laryngeal mechanism, M1
Article
A paired comparison technique in which fundamental frequency (330 Hz), phonemic category, /a/ and sound level (comfortable but equal) were held constant with register as the variable of primary interest was utilized with subjects who sang tone pairs varying in degree of audible‐register timbre difference or equalization. The paired tones were presented to a panel of 12 expert listeners for judgments of register identification and ratings of register equalization. The vowels produced were labeled by a trained phonetician. These judgments were compared with the acoustic records from a Sonograph modified to permit 24‐msec sections to be made at five positions within selected vibrato cycles: two peaks, two slopes, and a valley. The mean values for each partial and each subject were pooled in mean spectra for comparison of female chest and middle registers. The values for the paired tones judged to be “best equalized” were also compared across registers and with the spectra of the “least equalized” paired tones. It was found that chest register was characterized in general by greater energy in the partials above the third partial, especially partials 4 and 5, and that the middle register had a stronger fundamental.
Book
A vocal register is a series or range of consecutivefrequencies that can be produced with nearly identical voice quality.On the basis of research three fundamental registers can be definedand described: pulse, a low range of phonation; modal, a middle or"normal" range; and loft, a high range, described by some asfalsetto. These three principal vocal registers are defined on thebasis of four operational criteria. The acoustic characteristics ofthe registers are that each one occupies different ranges offundamental frequencies, reveals different magnitudes of vocalintensity, and has a different frequency composition. Perceptually,the three registers can be differentiated and identified on the basisof voice quality. Physiologcally, they differ because of vocal lengthor thickness and vibratory patterns. Aerodynamic characteristicsrelate to subglottic pressure, air flow, glottal resistance, andvoice intensity.
Article
The purpose of the present study was to investigate the vibratory patterns of the vocal folds during pulse register phonation.Glottal area–time functions were calculated from three high speed laryngeal films (4000 frames/second) obtained during phonation of the schwa vowel in pulse register by a normally hearing and speaking adult female. The results for the first film indicated that each of 35 consecutive vibratory cycles of the vocal folds consisted of a single opening/closing gesture followed by a lengthy closed phase. The analysis of the second film revealed that each of 33 consecutive vibratory cycles consisted of a double opening/closing vocal fold pattern, followed by a long closed phase. For the third film, the results indicated that each of 26 consecutive vibratory cycles of the vocal folds consisted of either a double or triple opening/closing gesture followed by a lengthy closed period. From these data, it appears that one of the physiological descriptors of pulse phonation is multiple, as well as single, vocal fold vibratory patterning.
Article
Singers can produce great varieties of vocal quality, pitch, and intensity, despite the fact that they have only one pair of vocal folds as their sound generators. Comprehensive study of the anatomy and physiology of the vocal folds and the muscles that control them reveals that humans exhibit remarkably complex control over the characteristics of their vocal folds. Alterations in the length, stiffness, shape, and other characteristics combine with changes in air flow, resonance, and other activities to permit such extraordinary diversity. The vocalis, cricothyroid, and lateral cricoarytenoid muscles are of particular importance. Additional interdisciplinary studies are needed to clarify further the mysteries of the human singing voice.