Electroglottographic Waveform Research Articles

Concurrent visualization and analysis of acoustic, flesh-point motion, and electroglottography signals during speech production

Over the years, technological advances have enabled speech researchers to directly track the skilled, sound-producing movements of the vocal tract, both intraoral and laryngeal articulators normally hidden from view (the tongue, velum, and glottis) and orofacial articulators directly visible on talkers’ faces (the lips and jaw). Despite these advances, however, no single instrument is capable of concurrently recording movements of all the articulators, which has impeded progress in characterizing inter-articulator control and coordination. To explore how inter-articulator coordination subserves linguistic structure, tools that co-register and temporally align different signals from different recording devices are necessary. This tutorial introduces optimal methods for studying the temporal coordination between laryngeal, intraoral, and orofacial articulators by combining various signals from electromagnetic articulography (EMA), electroglottography (EGG), and audio recordings, and displaying the time-aligned signals in the same analysis space. The multimodal data is processed using a set of MATLAB-based functions, which co-register and display positional and velocity trajectories of the lips, tongue, and jaw in tandem with the EGG waveform, F0 trajectory, and acoustic waveform and spectrogram. The coordination of laryngeal and supralaryngeal speech movements can then be measured and analyzed. [Work supported by an Emerging Research Grant from the Hearing Health Foundation.]

Normalized time-domain parameters for electroglottographic waveforms.

The electroglottographic waveform is of interest for characterizing phonation non-invasively. Existing parameterizations tend to give disparate results because they rely on somewhat arbitrary thresholds and/or contacting events. It is shown that neither are needed for formulating a normalized contact quotient and a normalized peak derivative. A heuristic combination of the two resolves also the ambiguity of a moderate contact quotient, with regard to vocal fold contacting being firm versus weak or absent. As preliminaries, schemes for electroglottography signal preconditioning and time-domain period detection are described that improve somewhat on similar methods. The algorithms are simple and compute quickly.

Effects of the Lung Volume on the Electroglottographic Waveform in Trained Female Singers

ObjectivesTo determine if in singing there is an effect of lung volume on the electroglottographic waveform, and if so, how it varies over the voice range. Study designEight trained female singers sang the tune “Frère Jacques” in 18 conditions: three phonetic contexts, three dynamic levels, and high or low lung volume. Conditions were randomized and replicated. MethodsThe audio and EGG signals were recorded in synchrony with signals tracking respiration and vertical larynx position. The first 10 Fourier descriptors of every EGG cycle were computed. These spectral data were clustered statistically, and the clusters were mapped by color into a voice range profile display, thus visualizing the EGG waveform changes under the influence of fo and SPL. The rank correlations and effect sizes of the relationships between relative lung volume and several adduction-related EGG wave shape metrics were similarly rendered on a color scale, in voice range profile-style ʻvoice maps.ʼ ResultsIn most subjects, EGG waveforms varied considerably over the voice range. Within subjects, reproducibility was high, not only across the replications, but also across the phonetic contexts. The EGG waveforms were quite individual, as was the nature of the EGG shape variation across the range. EGG metrics were significantly correlated to changes in lung volume, in parts of the range of the song, and in most subjects. However, the effect sizes of the relative lung volume were generally much smaller than the effects of fo and SPL, and the relationships always varied, even changing polarity from one part of the range to another. ConclusionsMost subjects exhibited small, reproducible effects of the relative lung volume on the EGG waveform. Some hypothesized influences of tracheal pull were seen, mostly at the lowest SPLs. The effects were however highly variable, both across the moderately wide fo-SPL range and across subjects. Different singers may be applying different techniques and compensatory behaviors with changing lung volume. The outcomes emphasize the importance of making observations over a substantial part of the voice range, and not only of phonations sustained at a few fundamental frequencies and sound levels.

Japanese macaque phonatory physiology.

Although the call repertoire and its communicative function are relatively well explored in Japanese macaques (Macaca fuscata), little empirical data are available on the physics and the physiology of this species' vocal production mechanism. Here, a 6 year old female Japanese macaque was trained to phonate under an operant conditioning paradigm. The resulting 'coo' calls and spontaneously uttered 'growl' and 'chirp' calls were recorded with sound pressure level (SPL) calibrated microphones and electroglottography (EGG), a non-invasive method for assessing the dynamics of phonation. A total of 448 calls were recorded, complemented by ex vivo recordings on an excised Japanese macaque larynx. In this novel multidimensional investigative paradigm, in vivo and ex vivo data were matched via comparable EGG waveforms. Subsequent analysis suggests that the vocal range (range of fundamental frequency and SPL) of the macaque was comparable to that of a 7-10 year old human, with the exception of low intensity chirps, the production of which may be facilitated by the species' vocal membranes. In coo calls, redundant control of fundamental frequency in relation to SPL was also comparable to that in humans. EGG data revealed that growls, coos and chirps were produced by distinct laryngeal vibratory mechanisms. EGG further suggested changes in the degree of vocal fold adduction in vivo, resulting in spectral variation within the emitted coo calls, ranging from 'breathy' (including aerodynamic noise components) to 'non-breathy'. This is again analogous to humans, corroborating the notion that phonation in humans and non-human primates is based on universal physical and physiological principles.

Modal and non-modal voice quality classification using acoustic and electroglottographic features.

The goal of this study was to investigate the performance of different feature types for voice quality classification using multiple classifiers. The study compared the COVAREP feature set; which included glottal source features, frequency warped cepstrum and harmonic model features; against the mel-frequency cepstral coefficients (MFCCs) computed from the acoustic voice signal, acoustic-based glottal inverse filtered (GIF) waveform, and electroglottographic (EGG) waveform. Our hypothesis was that MFCCs can capture the perceived voice quality from either of these three voice signals. Experiments were carried out on recordings from 28 participants with normal vocal status who were prompted to sustain vowels with modal and non-modal voice qualities. Recordings were rated by an expert listener using the Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V), and the ratings were transformed into a dichotomous label (presence or absence) for the prompted voice qualities of modal voice, breathiness, strain, and roughness. The classification was done using support vector machines, random forests, deep neural networks and Gaussian mixture model classifiers, which were built as speaker independent using a leave-one-speaker-out strategy. The best classification accuracy of 79.97% was achieved for the full COVAREP set. The harmonic model features were the best performing subset, with 78.47% accuracy, and the static+dynamic MFCCs scored at 74.52%. A closer analysis showed that MFCC and dynamic MFCC features were able to classify modal, breathy, and strained voice quality dimensions from the acoustic and GIF waveforms. Reduced classification performance was exhibited by the EGG waveform.

Voice quality variation over the course of the English utterance

The presence of atypical voice quality (e.g. breathy and creaky voice) can be used to diagnose voice disorders, but it is also clear that healthy English speakers’ voices vary as a function of phrasing. For example, ends of declarative utterances tend to be produced with creakier voice quality. Yet for speakers who have no diagnosed voice disorder, it is still unclear what factors affect voice variation over an utterance, and how consistent these variations are across speakers. This in turn limits diagnostic utility of perceived non-modal voice. Therefore, the goal of this project is to determine how the voice varies over the course of English utterances. We recorded electroglottographic (EGG) and audio waveforms of 20 male and female speakers of Californian English reading sentences of various lengths that were designed to avoid non-modal voice associated with certain segments. The EGG waveforms were analyzed for contact quotient and f0 at the beginning, middle, and end of the utterance. Discussion will focus on how voice quality differs by position in utterance, and the extent to which it is predictable from f0. We also discuss whether voice quality varies as a function of gender, utterance length, and speakers’ average voice quality.

A Measure of the Auditory-perceptual Quality of Strain from Electroglottographic Analysis of Continuous Dysphonic Speech: Application to Adductor Spasmodic Dysphonia

(1) To develop an automated algorithm to analyze electroglottographic (EGG) signal in continuous dysphonic speech, and (2) to identify EGG waveform parameters that correlate with the auditory-perceptual quality of strain in the speech of patients with adductor spasmodic dysphonia (ADSD). Software development with application in a prospective controlled study. EGG was recorded from 12 normal speakers and 12 subjects with ADSD reading excerpts from the Rainbow Passage. Data were processed by a new algorithm developed with the specific goal of analyzing continuous dysphonic speech. The contact quotient, pulse width, a new parameter peak skew, and various contact closing slope quotient and contact opening slope quotient measures were extracted. EGG parameters were compared between normal and ADSD speech. Within the ADSD group, intra-subject comparison was also made between perceptually strained syllables and unstrained syllables. The opening slope quotient SO7525 distinguished strained syllables from unstrained syllables in continuous speech within individual subjects with ADSD. The standard deviations, but not the means, of contact quotient, EGGW50, peak skew, and SO7525 were different between normal and ADSD speakers. The strain-stress pattern in continuous speech can be visualized as color gradients based on the variation of EGG parameter values. EGG parameters may provide a within-subject measure of vocal strain and serve as a marker for treatment response. The addition of EGG to multidimensional assessment may lead to improved characterization of the voice disturbance in ADSD.

Relationship Between the Electroglottographic Signal and Vocal Fold Contact Area

Electroglottography (EGG) is a widely used noninvasive method that purports to measure changes in relative vocal fold contact area (VFCA) during phonation. Despite its broad application, the putative direct relation between the EGG waveform and VFCA has to date only been formally tested in a single study, suggesting an approximately linear relationship. However, in that study, flow-induced vocal fold (VF) vibration was not investigated. A rigorous empirical evaluation of EGG as a measure of VFCA under proper physiological conditions is therefore still needed. Three red deer larynges were phonated in an excised hemilarynx preparation using a conducting glass plate. The time-varying contact between the VF and the glass plate was assessed by high-speed video recordings at 6000 fps, synchronized to the EGG signal. The average differences between the normalized [0, 1] VFCA and EGG waveforms for the three larynges were 0.180 (±0.156), 0.075 (±0.115), and 0.168 (±0.184) in the contacting phase and 0.159 (±0.112), -0.003 (±0.029), and 0.004 (±0.032) in the decontacting phase. Overall, there was a better agreement between VFCA and the EGG waveform in the decontacting phase than in the contacting phase. Disagreements may be caused by nonuniform tissue conductance properties, electrode placement, and electroglottograph hardware circuitry. Pending further research, the EGG waveform may be a reasonable first approximation to change in medial contact area between the VFs during phonation. However, any quantitative and statistical data derived from EGG should be interpreted cautiously, allowing for potential deviations from true VFCA.

Comparison of Tongue and Lip Trills with Phonation of the Sustained Vowel /<i>ε</i>/ Regarding the Periodicity of the Electroglottographic Waveform and the Amplitude of the Electroglottographic Signal

Aim: The aim is to compare the vocal fold vibration seen during lip and tongue trills with that seen during phonation of the sustained vowel /ε/, in terms of the periodicity of the EGG waveform and the amplitude of the EGG signal, in professional voice users. Study design: This was a quasi-experimental study. Methods: We used electroglottography (EGG) to compare the vocal fold vibration seen during tongue and lip trills with that seen during phonation of the sustained vowel /ε/, in terms of the EGG waveform periodicity and signal amplitude, in 10 classically trained, professional singers. The participants produced the sustained vowel /ε/ and performed tongue and lip trills at the same frequency and intensity. The periodicity of the waveform and the amplitude of the signal were visually analyzed by three blinded, experienced readers. To confirm the visual analysis results, we measured the jitter and shimmer of the signal and the frequency of variation in vocal fold vibration during the trill exercises. Results: The EGG waveform was classified as periodic for the sustained vowel phonation task and as quasi-periodic for the trill exercises, the vibration pattern repeating at approximately 24 Hz. Conclusion: The vibration of the vocal folds was modified according to the supraglottic movement in trills exercises.

A Preliminary Study for a Slantwise-Placed Electroglottography

Slantwise-placed electroglottography (EGG) was proposed for synchronous use with an ultrasound machine in a previous work. The objective was to confirm the feasibility of this slant EGG, differentiate it from conventional EGG, and give suggestions for its applications. A synchronized system composed of an EGG device, a high-speed video, a sound level meter, and a headset was established. The same phonation conditions were acquired by training subjects to reach the target phonation frequency and intensity. Electrode position was designed, and vocal fold vibration parameters were measured from EGG waveforms recorded in each electrode position. Comparison showed that the characteristic points identified in slant EGG waveforms were consistent with the vibration phase shown by high-speed video images. Phonation frequency measured from slant EGG was highly accurate. EGG amplitude nonlinearly decreased with the increase of electrode distance. Compared with conventional EGG parameters, velocity ratio and glottal closed quotient measured from slant EGG were accurate when electrodes were placed symmetrically and the electrode distance was within a proper length. Slant EGG was proved feasible and can be considered as a useful tool to obtain comprehensive information in investigating the invivo vocal fold dynamics when synchronized with other detecting equipment.

The Effect of Gender on Measures of Electroglottographic Contact Quotient

The purpose of this study was to reevaluate the possible effect of gender on measures of electroglottographic contact quotient (CQ). In addition to criterion threshold methods frequently used for the computation of CQ's via electroglottography (EGG), derivative based methods of computing the EGG CQ that have not been previously evaluated in regard to their effectiveness in discriminating between male versus female EGG waveforms were also examined. Causal, mixed design. EGG waveforms were recorded during sustained vowel production from normal voice men and women (total N=50). Measures of EGG CQ were calculated using eight criterion thresholds (25-60% in 5% increments). In addition, CQs were computed using the EGG derivative (DEGG) as well as via two hybrid methods that combine thresholds with measures from the DEGG. Only CQs calculated via the DEGG showed a statistically significant difference in the direction anticipated via previous research (ie, men>women). In addition, the DEGG was the only method to show statistically significant differences in CQ useful in discriminating between "knee"-shaped opening phases observed in 76% of the men versus no "knee" opening phases observed in 64% of the women. The DEGG may provide advantages over other methods of EGG CQ analysis that are useful in revealing possible gender differences in vocal physiology. Further studies that examine the relationship between EGG, DEGG, and visualization of vocal fold dynamics via high-speed video may provide further insight into the possible differences between adult male versus female contacting and decontacting points during comfortable pitch and loudness phonation.

Laryngeal vibratory behavior in traditional Noh singing

The phonation of Noh, a traditional Japanese style of singing, was investigated using electroglottographic and acoustical analyses. The dynamics of the laryngeal vibratory behaviors were analyzed for the singing voice of the Noh play compared with natural speech based on the electroglottography (EGG) parameters, EGG waveform, spectrum and spectrogram. The result shows that Noh singing is characterized by low OQegg and high SQegg. Three types of phonations are used in the singing with pressed, vocal-ventricular mode (VVM), and growl voices. It was hypothesized that the period doubling observed in the EGG signal was reflective of VVM, which was caused by the phase difference in the vocal and ventricular fold oscillations, while the damped peak amplitude in every other cycle in the EGG signal was the result of the oscillations of the aryepiglottic folds at a frequency of half of the fundamental frequency. Subharmonics generated by the supraglottal oscillations add unique timbre to the sounds. The results suggest that the combination of phonation types is the key factor in generating their peculiar voice qualities.

Physiological and acoustic characteristics of the female music theater voice

Three Music Theater vocal qualities ("chesty belt," "twangy belt," and "legit") were investigated in female singers at their overlap range, between F#4-D5 (~370-600 Hz). Six experienced Music Theater singers performed each quality on two different vowels ([e], [ɔ]). Audio and electroglottographic (EGG) signals were recorded as well as the vocal tract impedance. In chesty belt and twangy belt, singers systematically tuned the frequency of their first vocal tract resonance (R1) to the second harmonic (2f(0)) up to C5. R1 remained lower in frequency for the legit quality. No tuning of the second vocal tract resonance (R2) was observed in any of these qualities although R2 frequency was significantly higher in both belt qualities than in legit. Closed quotient, degree of symmetry of the EGG waveform, sound pressure level (SPL) and the energy of the spectrum above 1 kHz were significantly greater in chesty belt than in legit but not significantly different between chesty belt and twangy belt qualities. A fourth quality ("mix") was explored in three singers. Different production strategies were observed for each singer, with values of spectral, glottal and vocal tract descriptors found in between those measured for legit and chesty belt qualities.

Correlates of the Belt Voice: A Broader Examination

The purpose of the present study was to further clarify a definition of belting in physiological and acoustic terms. Twenty professional female singers produced triads ascending by halftone steps from C₄ to C₅ and sang the last four measures of a popular Broadway ballad in both legit and belt styles. Vocal source differences between belt and legit singing modes were assessed with respect to closed quotient (CQ) and speed quotient (SQ) computed from the electroglottographic waveform during productions of /ma/ triads covering approximately one octave. Differences in supraglottal configuration were evaluated in terms of the ratio (dB) of the amplitudes of the first and second harmonics of /I/ sustained at the pitch A₅(b), as noted in the long-term average spectrum. CQ did not vary significantly as a function of the singing type. SQ, however, decreased monotonically in association with increasing pitch level in both legit and belt singing styles. Belt mode phonations had higher SQ's than those of legit mode phonations at the same pitch. A repeated-measures analysis of variance demonstrated that the data differences associated with both pitch level and with singing mode were significant beyond P < 0.01. Analysis of the relative magnitude of the first and second harmonics showed that the direction of the harmonic ratio varied between styles. In legit singing style, H₁ was always stronger than H₂, but this unidirectionality of amplitude was not characteristic of belt singing, for which the second harmonic had greater amplitude in 25% of the cases. Implications of these results are discussed with regard to a new definition of belting.

Using Electroglottographic Real-Time Feedback to Control Posterior Glottal Adduction during Phonation

The goal of this pilot study was to determine whether the ability to change the degree of posterior glottal adduction (PGA) during phonation can be acquired more easily with the aid of electroglottographic (EGG) real-time feedback. The subject was a 37-year-old untrained female with habitually breathy voice. Before the experiment, she participated in one voice coaching session where exercises for increasing PGA were explained and executed. During the experiment, phonation has been monitored simultaneously with videostroboscopy, electroglottography, and audio recording. While phonating, the subject saw amplitude and period normalized EGG waveform representing one glottal cycle consecutively changing over time. The assignment was to increase the width of the EGG waveform during phonation. Laryngeal imaging revealed a posterior glottal chink during habitual phonation. The subject could only introduce intentional changes into the EGG waveform after its relevance had been explained, and after recapitulation of the exercises of the voice coaching session: An increase of the EGG waveform width coincided with the increase of high-frequency partials and an increase of PGA. For pitches B3 and B4, full glottal closure could be achieved. At G5, a reduction of the posterior glottal chink occurred. The findings of this study suggest that the skill to control the degree of PGA can be acquired, and that EGG real-time feedback can be a crucial element in optimizing the process of skill acquisition, but only if (1) the context and nature of the feedback is explained and (2) proper instructions are provided. The EGG contact quotient might not be sensitive to changes of PGA in falsetto phonation.

Laryngeal Electroglottography as a Predictor of Laryngeal Electromyography

We evaluate a group of patients who have mobile vocal folds, but have flexible laryngeal examinations suspicious for mild vocal fold paresis. The purpose of this retrospective study is to evaluate if electroglottography (EGG) can predict the probability of having an abnormal laryngeal electromyography (LEMG). Charts of patients evaluated and suspected of having mild vocal fold paresis between August 1, 2004 and March 30, 2006 were reviewed. We compared average EGG contact quotients (CQs), average absolute difference of recorded CQ from normal (|CQ-50%|), and EGG waveforms between patients with normal and abnormal LEMG. Waveforms were evaluated in blinded fashion. Statistical analysis was performed using chi square and t test analyses. One hundred and sixteen patients received both flexible laryngoscopy and LEMG as part of their evaluation. Forty-eight patients (41%) had confirmed paresis by LEMG and 68 patients (59%) had normal LEMGs. Only 9.1% of patients with a normal EGG waveform had an abnormal LEMG. In contrast, 40.4% of patients with an abnormal EGG waveform had abnormal LEMGs. The negative predictive value of a normal EGG waveform for an abnormal LEMG was 90.9%. No significant differences were identified between patients with normal versus abnormal LEMG in terms of average CQ (47.8% vs 46.4%) or |CQ-50%| (6.2 vs 5.6). Patients with a normal EGG waveform are unlikely to have an abnormal LEMG.

On voicing activity under the control of emotion and loudness

For a parametric study of interplay between loudness control and emotional modulation in voicing activity, electroglottography (EGG) data collected from two male and two female subjects are examined. The subjects read emotionally neutral sentences in a self-controlled manner with four different emotional states (neutral, angry, sad, happy) and three levels of loudness (soft, normal, loud). The analysis focused on the timing and shape-related EGG waveform parameters of the vowel /a/ in the data as a function of emotional state and loudness level. Specifically, open quotient (OQ), speed quotient (SQ), and noise-to-harmonic ratio (NHR) of EGG waveform are investigated. Pitch, root-mean-square (RMS) energy, and duration of speech waveform are also analyzed. Despite inter-subject differences, some general tendencies of the EGG parameters can be observed. It is found that angry emotion shows the lowest OQ in a given loudness level, and happy emotion exhibits significantly higher OQ than angry emotion. SQ mainly varies along the loudness dimension, smaller SQ in louder voice. NHR is the highest in soft voice and shows no clear emotion-dependent pattern. The results suggest that OQ is the main EGG parameter that is controlled by speakers for emotion expression, in addition to its role in loudness control.

An integrated tool for the diagnosis of voice disorders

A PC-based integrated aid tool has been developed for the analysis and screening of pathological voices. With it the user can simultaneously record speech, electroglottographic (EGG), and videoendoscopic signals, and synchronously edit them to select the most significant segments. These multimedia data are stored on a relational database, together with a patient's personal information, anamnesis, diagnosis, visits, explorations and any other comment the specialist may wish to include. The speech and EGG waveforms are analysed by means of temporal representations and the quantitative measurements of parameters such as spectrograms, frequency and amplitude perturbation measurements, harmonic energy, noise, etc. are calculated using digital signal processing techniques, giving an idea of the degree of hoarseness and quality of the voice register. Within this framework, the system uses a standard protocol to evaluate and build complete databases of voice disorders. The target users of this system are speech and language therapists and ear nose and throat (ENT) clinicians. The application can be easily configured to cover the needs of both groups of professionals. The software has a user-friendly Windows style interface. The PC should be equipped with standard sound and video capture cards. Signals are captured using common transducers: a microphone, an electroglottograph and a fiberscope or telelaryngoscope. The clinical usefulness of the system is addressed in a comprehensive evaluation section.

Gender differences in vocal fold contact computed from electroglottographic signals: the influence of measurement criteria.

EGGW is a phonatory parameter that can be derived from electroglottographic (EGG) signals and used to infer the relative degree of vocal fold contact. Vocal fold models predict that men will exhibit medial bulging of their vocal folds during phonation but women will not. These models lead us to expect gender differences in the magnitude of EGGW. Nevertheless, significant gender differences in EGGW for adults with normal voices have not been documented in previous studies when EGGW was computed from criterion lines placed at 25%-40% of the amplitude of the uninverted EGG wave form. We hypothesized that EGGW would better reflect gender differences in vocal fold adductory patterns if EGGW was computed from portions of the wave form that were associated with more vocal fold contact. EGGW was measured for seven men and seven women with normal voices. When EGGW was computed from segments of the wave form that were associated with relatively greater vocal fold contact (i.e., using criterion levels of > or = 55%), findings were consistent with the gender-specific adductory patterns that have been proposed from vocal fold models. Guidelines for appropriate placement of criterion lines when computing EGGW are discussed.

Glottographic phase difference in recurrent nerve paralysis.

Phase measures with simultaneously recorded electroglottography (EGG) and photoglottography (PGG) signals have been studied in canine models and found to be sensitive to the effect of recurrent laryngeal nerve (RLN) paralysis on vocal fold vibration. This study examined the usefulness of this type of measure in clinical application. The combined glottographic signals were obtained from 5 men with a diagnosis of unilateral RLN paralysis and 5 age-matched controls. In the patient group, EGG waveforms were found to have a consistent phase delay in relation to PGG. A measure of the overall glottographic phase difference (GPD) was found to significantly distinguish the control group (mean GPD = 0.2371 ms) from the patient group (mean GPD = -0.2765 ms). A validity analysis performed on 19 subjects with or without unilateral RLN paralysis confirmed that the GPD test had a relatively high efficiency (91.7%) in detecting unilateral RLN paralysis in the male population.