Healthy Voices Research Articles

Automatic Classification of Synthetic Voices for Voice Banking Using Objective Measures

Speech is the most common way of communication among humans. People who cannot communicate through speech due to partial of total loss of the voice can benefit from Alternative and Augmentative Communication devices and Text to Speech technology. One problem of using these technologies is that the included synthetic voices might be impersonal and badly adapted to the user in terms of age, accent or even gender. In this context, the use of synthetic voices from voice banking systems is an attractive alternative. New voices can be obtained applying adaptation techniques using recordings from people with healthy voice (donors) or from the user himself/herself before losing his/her own voice. In this way, the goal is to offer a wide voice catalog to potential users. However, as there is no control over the recording or the adaptation processes, some method to control the final quality of the voice is needed. We present the work developed to automatically select the best synthetic voices using a set of objective measures and a subjective Mean Opinion Score evaluation. A prediction algorithm of the MOS has been build which correlates similarly to the most correlated individual measure.

When I Think of Home: Black Families Supporting Their Children During the COVID-19 Pandemic.

Since the onset of the COVID-19 pandemic, economic crisis, and persistent systemic and structural racism have plagued Black communities. The continued physical and symbolic violence and murders of Black bodies are undeniable. As White institutions, schools are definite contributors to this brutality as they center the culture and realities of White children while ignoring or denigrating Black children. This is even evident in the undermining of Black families' efforts to prepare their children to face the inequities and injustices they experience in the U.S. In this article, we discuss Black families' engagement in their children's education amid threats through racial socialization research aimed at developing and validating Black children's perspectives, experiences, and realities in Black identity to promote their positive social-emotional and psychological development. Black families must know how to cultivate their child's healthy self-identity, voice, and agency, along with academic achievement. Schools should learn from these practices. Schools that choose to ignore these concepts will continue contributing to trauma and violence against Black children and maintain deficit-oriented views. The article includes examples and implications for teaching and supporting the well-being of Black children, and concludes with practical ideas that educators can learn from and integrate into their practices.

Outcomes of Gender Affirming Voice Training: A Comparison of Hybrid and Individual Training Modules.

Gender affirming voice training supports individuals as they explore and develop a healthy voice that feels congruent with their gender and sense of self. Although gender affirming voice training is increasingly available, there is no research comparing individual versus hybrid training outcomes in the existing literature. A retrospective chart review was performed on trans women and/or nonbinary individuals who attended either individual or hybrid (group and individual training) gender affirming voice and communication training at St. Michael's Hospital in Toronto between 2012 and 2019. Participant demographics, self-reported goal achievement, attendance, pre/post-Trans Woman Voice Questionnaire (TWVQ) scores, and pre/postspeaking fundamental frequency (SFF) were compared between individual and hybrid training participants. Multiple logistic regression analysis was performed to determine factors that could predict self-reported goal achievement. Sixty-two trans women and/or nonbinary individuals who attended either individual (n = 43) or hybrid (n = 19) gender affirming voice and communication training were included in the study. Hybrid participants had a significantly higher goal achievement rate and therapy adherence rate compared with individual participants (74% vs. 21%; 84% vs. 28%, respectively). Both training models resulted in significant elevation of SFF and reduction of TWVQ scores. Inflection range did not change significantly for either training formats. Multiple logistic regression showed that hybrid training was a positive predictor for goal achievement (odds ratio = 2.2, 95% confidence interval [0.3, 4.1]). Although clients in individual and hybrid gender affirming voice and communication training achieved significant pitch elevation and lower TWVQ scores, hybrid participants demonstrated significantly higher goal achievement and therapy adherence. This study suggests that hybrid training may provide a valuable support system that improves therapy success in this client population.

Artificial Intelligence Techniques for the Non-invasive Detection of COVID-19 Through the Analysis of Voice Signals.

Healthcare sensors represent a valid and non-invasive instrument to capture and analyse physiological data. Several vital signals, such as voice signals, can be acquired anytime and anywhere, achieved with the least possible discomfort to the patient thanks to the development of increasingly advanced devices. The integration of sensors with artificial intelligence techniques contributes to the realization of faster and easier solutions aimed at improving early diagnosis, personalized treatment, remote patient monitoring and better decision making, all tasks vital in a critical situation such as that of the COVID-19 pandemic. This paper presents a study about the possibility to support the early and non-invasive detection of COVID-19 through the analysis of voice signals by means of the main machine learning algorithms. If demonstrated, this detection capacity could be embedded in a powerful mobile screening application. To perform this important study, the Coswara dataset is considered. The aim of this investigation is not only to evaluate which machine learning technique best distinguishes a healthy voice from a pathological one, but also to identify which vowel sound is most seriously affected by COVID-19 and is, therefore, most reliable in detecting the pathology. The results show that Random Forest is the technique that classifies most accurately healthy and pathological voices. Moreover, the evaluation of the vowel /e/ allows the detection of the effects of COVID-19 on voice quality with a better accuracy than the other vowels.

Classification of Parkinson disease based on analysis and synthesis of voice signal

The most important application of voice profiling is pathological voice detection. Parkinson's disease is a chronic neurological degenerative disease affecting the central nervous system responsible for essentially progressive evolution movement disorders. 70% to 90% of Parkinson’s disease (PD) patients show an affected voice. This paper proposes a methodology for PD based on acoustic, glottal, physical, and electrical parameters. The results show that the acoustic parameter is more important in the case of Parkinson’s disease as compared to glottal and physical parameters. The authors achieved 97.2% accuracy to differentiate Parkinson and healthy voice using jitter to pitch ratio proposed algorithm. The Authors also proposed an algorithm of poles calculation of the vocal tract to find formants of the vocal tract. Further, formants are used for finding the transfer function of vocal tract filter. In the end, the authors suggested parameters of the electrical vocal tract model are also changed in the case of PD voices.

Classification of Parkinson disease based on analysis and synthesis of voice signal

The most important application of voice profiling is pathological voice detection. Parkinson's disease is a chronic neurological degenerative disease affecting the central nervous system responsible for essentially progressive evolution movement disorders. 70% to 90% of Parkinson’s disease (PD) patients show an affected voice. This paper proposes a methodology for PD based on acoustic, glottal, physical, and electrical parameters. The results show that the acoustic parameter is more important in the case of Parkinson’s disease as compared to glottal and physical parameters. The authors achieved 97.2% accuracy to differentiate Parkinson and healthy voice using jitter to pitch ratio proposed algorithm. The Authors also proposed an algorithm of poles calculation of the vocal tract to find formants of the vocal tract. Further, formants are used for finding the transfer function of vocal tract filter. In the end, the authors suggested parameters of the electrical vocal tract model are also changed in the case of PD voices.

Intelligibility of dysphonic speech in auralized classrooms.

Voice disorders can reduce the speech intelligibility of affected speakers. This study evaluated the effect of noise, voice disorders, and room acoustics on vowel intelligibility, listening easiness, and the listener's reaction time. Three adult females with dysphonia and three adult females with normal voice quality recorded a series of nine vowels of American English in /h/-V-/d/ format (e.g., "had"). The recordings were convolved with two oral-binaural impulse responses acquired from measurements in two classrooms with 0.4 and 3.1 s of reverberation time, respectively. The stimuli were presented in a forced-choice format to 29 college students. The intelligibility and the listening easiness were significantly higher in quiet than in noisy conditions, when the speakers had normal voice quality compared to a dysphonic voice, and in low reverberated environments compared to high reverberated environments. The response time of the listener was significantly longer for speech presented in noisy conditions compared to quiet conditions and when the voice was dysphonic compared with healthy voice quality.

Gammatone spectral latitude features extraction for pathological voice detection and classification

To improve the performance of pathological voice detection and classification, gammatone spectral latitude (GTSL) features were proposed. GTSL features are inspired by the nonlinear phenomena produced from the human phonation, presenting explicit physiological meaning. The features combine with human auditory perception characteristics. GTSL features quantify the turbulent noise by the nonlinear compression of peak value and dynamic range of the spectrums in each frequency channel. For pathological voice detection, gammatone spectral latitude (GTSL) features fitted better with traditional machine learning algorithms than traditional nonlinear features and gammatone ceptral coefficients (GTCCs). In the classification between healthy, neuromuscular and structural voices, the proposed features achieved average accuracy of 99.6% in the Massachusetts Eye and Ear Infirmary (MEEI) database, which is 35.6% higher than other gammatone features. The accuracies in other database, Saarbruecken Voice Database (SVD) and Hospital Universitario Prłncipe de Asturias (HUPA), were 89.9% and 97.4% respectively. The experimental results indicate that, GTSL features can provide objective evaluation of voice diseases with low computational complexity and database dependency.

Classification of Parkinson Disease Based on Analysis and Synthesis of Voice Signal

The most important application of voice profiling is pathological voice detection. Parkinson's disease is a chronic neurological degenerative disease affecting the central nervous system responsible for essentially progressive evolution movement disorders. 70% to 90% of Parkinson’s disease (PD) patients show an affected voice. This paper proposes a methodology for PD based on acoustic, glottal, physical, and electrical parameters. The results show that the acoustic parameter is more important in the case of Parkinson’s disease as compared to glottal and physical parameters. The authors achieved 97.2% accuracy to differentiate Parkinson and healthy voice using jitter to pitch ratio proposed algorithm. The Authors also proposed an algorithm of poles calculation of the vocal tract to find formants of the vocal tract. Further, formants are used for finding the transfer function of vocal tract filter. In the end, the authors suggested parameters of the electrical vocal tract model are also changed in the case of PD voices.

Deep connected attention (DCA) ResNet for robust voice pathology detection and classification

The automatic diagnosis method based on speech signal analysis is able to realize the detection and classification of pathological voices. It plays an important role in the early diagnosis and auxiliary treatment of voice pathology, which effectively relief the discomfort of patients and reduce the workload of doctors. Therefore, the automatic diagnosis method based on speech signal analysis is of great research value. Meanwhile, high accuracy, high precision and stability are the pursuit goals. In this paper, a novel computer-aided assessment based on speech signal analysis for pathological voice classification (CS-PVC) system is proposed. This model focuses on the areas with large differences between different pathological voices and healthy voices, while ignore the negative impact of insignificant information on the performance of the model. Two databases were used in the experiments, one is the Saarbruecken Voice database (SVD), and the other is the self-built Shenzhen People’s Hospital voice database (SZUPD). The pathological voice detection accuracy of the proposed system on the above two databases are 81.6% and 82.2% respectively. The experimental results show that the proposed framework is not data-dependence. In other words, it has the potential to be universally applicable in medical framework in the future.

Influence of Reduced Saliva Production on Phonation in Patients With Ectodermal Dysplasia

Patients with ectodermal dysplasia (ED) suffer from an inherited disorder in the development of the ectodermal structures. Besides the main symptoms, i.e. significantly reduced formation/expression of teeth, hair and sweat glands, a decreased saliva production is objectively accounted. In addition to difficulties with chewing/swallowing, ED patients frequently report on the subjective impression of rough and hoarse voices. A correlation between the reduced production of saliva and an affliction of the voice has not yet been investigated objectively for this rare disease. Following an established measurement protocol, a study has been conducted on 31 patients with ED and 47 controls (no ED, healthy voice). Additionally, the vocal fold oscillations were recorded by high-speed videoendoscopy (HSV@4kHz). The glottal area waveform was determined by segmentation and objective glottal dynamic parameters were calculated. The generated acoustic signal was evaluated by objective and subjective measures. The individual impairment was documented by a standardized questionnaire (VHI). Additionally, the amount of generated saliva was measured for a defined period of time. ED patients displayed a significantly reduced saliva production compared to the control group. Furthermore, the auditory-perceptual evaluation yielded significantly higher ratings for breathiness and hoarseness in the voices of male ED patients compared to male controls. The majority of male ED patients (67%) indicated at least minor impairment in the self-evaluation. Objective acoustic measures like Jitter and Shimmer confirmed the decreased acoustic quality in male ED patients, whereas none of the investigated HSV parameters showed significant differences between the test groups. Statistical analysis did not confirm a statistically significant correlation between reduced voice quality and amount of saliva. An objective impairment of the acoustic outcome was demonstrated for male ED patients. However, the vocal folds dynamics in HSV recordings seem unaffected.

Acoustic Variability in the Healthy Female Voice Within and Across Days: How Much and Why?

Purpose The aims of this study were (1) to quantify variability in voice production (as measured acoustically) within and across consecutive days in vocally healthy female speakers, (2) to identify which acoustic measures are sensitive to this variability, and (3) to identify participant characteristics related to such voice variability. Method Participants included 45 young women with normal voices who were stratified by age, specifically 18-23, 24-29, and 30-35 years. Following an initial acoustic and auditory-perceptual voice assessment, participants performed standardized field voice recordings 3 times daily across a 7-day period. Acoustic analyses involved 32 cepstral-, spectral-, and time-based measures of connected speech and sustained vowels. Relationships among acoustic data and select demographic, health, and lifestyle (i.e., participant-based) factors were also examined. Results Significant time-of-day effects were observed for acoustic analyses within speakers (p < .05), with voices generally being worse in the morning. No significant differences were observed across consecutive days. Variations in voice production were associated with several participant factors, including improved voice with increased voice use; self-perceived poor voice function, minimal or no alcohol consumption, and extroverted personality; and worse voice with regular or current menstruation, depression, and anxiety. Conclusions This acoustic study provides essential information regarding the nature and extent to which healthy voices vary throughout the day and week. Participant-based factors that were associated with improved voice over time included increased voice use, self-perceived poor voice function, minimal or no alcohol consumption, and extroverted personality. Factors associated with worse voice production over time included regular or current menstruation, and depression and anxiety.

A Cross-sectional Study of Perceptual and Acoustic Voice Characteristics in Healthy Aging

The human voice qualitatively changes across the lifespan. Although some of these vocal changes may be pathologic, other changes likely reflect natural physiological aging. Normative data for voice characteristics in healthy aging is limited and disparate studies have used a range of different acoustic features, some of which are implicated in pathologic voice changes. We examined the perceptual and acoustic features that predict healthy aging. Participants (N = 150) aged between 50 and 92 years performed a sustained vowel task. Acoustic features were measured using the Multi-Dimensional Voice Program and the Analysis of Dysphonia in Speech and Voice. We used forward and backward variable elimination techniques based on the Bayesian information criterion and linear regression to assess which of these acoustic features predict age and perceptual features. Hearing thresholds were determined using pure-tone audiometry tests at frequencies 250 Hz, 500 Hz, 1000 Hz, 2000 Hz, and 4000 Hz. We further explored potential relationships between these acoustic features and clinical assessments of voice quality using the Consensus Auditory-Perceptual Evaluation of Voice. Chronological age was significantly predicted by greater voice turbulence, variability of cepstral fundamental frequency, low relative to high spectral energy, and cepstral intensity. When controlling for hearing loss, age was significantly predicted by amplitude perturbations and cepstral intensity. Clinical assessments of voice indicated perceptual characteristics of speech were predicted by different acoustic features. For example, breathiness was predicted by the soft phonation index, mean cepstral peak prominence, mean low-high spectral ratio, and mean cepstral intensity. Findings suggest that acoustic features that predict healthy aging are different than those previously reported for the pathologic voice. We propose a model of healthy and pathologic voice development in which voice characteristics are mediated by the inability to monitor vocal productions associated with age-related hearing loss. This normative data of healthy vocal aging may assist in separating voice pathologies from healthy aging.

Acoustic Voice Analysis of Normal and Pathological Voices in Indian Population Using Praat Software.

Acoustic voice analysis is still a valuable technique which enables voice clinicians to compare voices to differentiate them into normal and abnormal. The present study was undertaken to standardize acoustic voice parameters in normal healthy adult individuals and gender comparison among them and also acoustic voice analysis of pathological voices and it's comparison with normal healthy voices. Voice samples of vowels /a/, /i/ and /u/ of 80 normal healthy adults (males = 40, females = 40) of control group and 40 patients with dysphonic voice of case group collected and acoustic voice parameters were extracted by using Praat software. There were statistically significant higher values of fundamental frequency (F0) in females, while jitter local (%), shimmer local (%) and harmonic to noise ratio (HNR) had no gender differences in normal healthy voices. Pathological voices of case group subjects with laryngeal pathologies had statistically significant higher values of jitter local (%), shimmer local (%) and lower values of HNR as compare to normal healthy voices of control group. Objective voice analysis by using Praat software is convenient, reliable and cost effective method. This study establishes normative acoustic voice parameters in normal healthy adults. There are no gender differences in adult healthy voices except fundamental frequency (F0), which is higher in females. Patients who are with dysphonic voices due to laryngeal pathologies had altered values of acoustic parameters compared to normophonic adults and clinicians can precisely differentiate pathological voices from normophonics.

Investigating the cepstral acoustic characteristics of voice in healthy children

ObjectivesThis study aimed to determine the cepstral acoustic parameters that vary depending on age and gender in vocally healthy children, and to establish normative data for cepstral analysis. BackgroundCepstral measurements are among the strongest predictors of auditory-perceptual evaluation of voice and differentiate between healthy and dysphonic voices. More spesificially, ceptral peak prominence is accepted to be as a strong acoustic predictor of breathiness and overall severity of dysphonia. Cepstral measures determine voice quality reliably not only in sustained vowel samples but also in running speech samples. Determining the parameters related to the acoustic profile of children with normal voices can lead us to a better understanding of the effect of changes in the larynx and vocal fold structure during growth and development. There is a limited number of norm studies examining the cepstral acoustic properties of pediatric voice. Determining norm-specific values and clinical guidelines of cepstral acoustics according to the age and gender in vocally healthy children are utmost important. MethodsA total of 160 vocally healthy children were divided into the following four age groups: Group-I included children within the age range of 4–7 years, Group-II included 7–11 years, Group-III 11–14 years, and Group-IV included children within the age range of 14–18 years. An equal number of male and female participants were assigned to each group. PENTAX Medical CSL Model 4500 was used for recording all tasks. For acoustic analysis, Multi-Dimensional Voice Program and Analysis of Dysphonia in Speech and Voice were used. ResultsCepstral Peak Prominence (CPP), Cepstral Peak Prominence Standard Deviation (CPP SD), and Low-To-High Spectral Ratio (L/H Ratio) increased with age. It is found that the CPP parameter of all-voiced sentences and nasal-weighted sentences increased with age in boys, while no significant pattern was observed in any sample for girls. For L/H ratio, it can be said that there is a general increase with age in all speech samples, except for the vowel-weighted and voiceless plosive sentence samples, evident especially in the group above the age of 15 years. This study concluded that the CPP SD parameter in the vowel-weighted sentences increased with age in boys. It was also noticed in this study that CPP F0 standard deviation (SD) intervals were narrower in vowel-weighted, easy onset, and voiceless plosive sentence samples than in all-voiced, hard glottal attack and nasal-weighted sentence samples. ConclusionThis study established cepstral acoustic normative values for a wide age range of the pediatric population. It is thought that age and gender specific cepstral acoustic findings presented in this study contributed to the related literature. In addition, to our knowledge, this is the first study that provides a normative cepstral acoustic database of the CAPE-V/Turkish sentences in the pediatric population.

Normative Voice Range Profile of the Young Female Voice

Normative data are important in the clinical setting of Speech and Language Pathology. The purpose of this study was to develop a normative reference dataset of voice range profiles from young females. Descriptive study including a prospective collection of voice range profile data. Voice range profile recordings from 39 females with healthy voices aged 18 to 28 years were conducted. Seven voice range profile variables were analyzed: minimum and maximum fundamental frequency and intensity, semitone and intensity ranges, and voice range profile area. Descriptive statistical methods were applied. An age-specific voice range profile normative dataset was established. The mean values and standard deviations were as follows: semitone range 34.7 ± 3.9 ST, minimum fundamental frequency 143.6 ± 21.7 hertz, maximum fundamental frequency 1063.5 ± 160 hertz, intensity range 65.6 ± 5.0 dB, minimum intensity 43.2 ± 2.5 dB SPL, maximum SPL 108.9 ± 5.1 dB SPL, and voice range profile area 1346 ± 222 cells. A normative dataset usable for optimization of future voice assessments has been established. It may especially benefit evaluation and treatment planning for younger females suffering from vocal fold nodules.

How to Maintain Your Healthy Voice

How to Maintain Your Healthy Voice

Changes in Relative Fundamental Frequency Under Increased Cognitive Load in Individuals With Healthy Voices.

Purpose The purpose of this study was to determine the effect of cognitive load on relative fundamental frequency (RFF) in individuals with healthy voices. Method Twenty adults with healthy voices read sentences under different cognitive load conditions. Each sentence contained color terms printed in colored ink, creating an embedded Stroop task. Participants read the ink color in which a word was printed, rather than the color term itself. Sentences with mismatched ink colors and printed words constituted an increased cognitive load. RFF, an acoustic correlate of laryngeal tension, was calculated for the 10 voicing cycles preceding (i.e., offset) and following (i.e., onset) voiceless consonants. Repeated measures analyses of variance were constructed to assess the effects of RFF cycle, cognitive load, and their interaction on mean RFF offset and onset. Results There was a significant effect of cognitive load condition on RFF offset. There was no significant effect of condition on RFF onset nor significant interaction between cycle and condition on RFF onset or offset values. Conclusion Reduced mean RFF offset may indicate an increase in laryngeal muscle tension during a cognitively demanding task.

Aeroacoustic Sound Source Characterization of the Human Voice Production-Perturbed Convective Wave Equation

The flow-induced sound sources of human voice production are investigated based on a validated voice model. This analysis is performed using a hybrid aeroacoustic workflow based on the perturbed convective wave equation. In the first step, the validated 3D incompressible turbulent flow simulation is computed by the finite volume method using STARCCM+. In a second step, the aeroacoustic sources are evaluated and studied in detail. The formulation of the sound sources is compared to the simplification (neglecting the convective sources) systematically using time-domain and Fourier-space analysis. Additionally, the wave equation is solved with the finite element solver openCFS to obtain the 3D sound field in the acoustic far-field. During the detailed effect analysis, the far-field sound spectra are compared quantitatively, and the flow-induced sound sources are visualized within the larynx. In this contribution, it is shown that the convective part of the sources dominates locally near the vocal folds (VFs) while the local time derivative of the incompressible pressure is distributed in the whole supra-glottal area. Although the maximum amplitude of the time derivative is lower, the integral contribution dominates the sound spectrum. As a by-product of the detailed perturbed convective wave equation source study, we show that the convective source term can be neglected since it only reduces the validation error by 0.6%. Neglecting the convective part reduces the algorithmic complexity of the aeroacoustic source computation of the perturbed convective wave equation and the stored flow data. From the source visualization, we learned how the VF motion transforms into specific characteristics of the aeroacoustic sources. We found that if the VFs are fully closing, the aeroacoustic source terms yield the highest dynamical range. If the VFs are not fully closing, VFs motion does not provide as much source energy to the flow-induced sound sources as in the case of a healthy voice. As a consequence of not fully closing VFs, the cyclic pulsating velocity jet is not cut off entirely and therefore turbulent structures are permanently present inside the supraglottal region. These turbulent structures increase the broadband component of the voice signal, which supports research results of previous studies regarding glottis closure and insufficient voice production.

The Influence of Voice Training on Vocal Learner's Objective Acoustic Voice Components

Acoustic parameters of voice were studied in music majors throughout 18 months of training to understand the influence of voice training on voice. Twenty-three students from Xiamen Music School between 12 and 15 years old were enrolled. Acoustic examination was performed three times- every 6 months for 18 months. Various traditional acoustic parameters were measured, including dysphonia severity index (DSI), jitter, and D-value of vocal range. Nonlinear dynamic measures were also measured, including diffusive chaos to construct voice type component profiles (VTCPs), spectrum convergence ratio, and nonlinear energy difference ratio. The results were analyzed by multivariate analysis of variance. Over the study duration, there was an improvement of DSI (P=0.002), and D-value of vocal range (P=0.000). Among nonlinear parameters, only voice type component data demonstrated significant changes during the study duration. Both Voice Type Component 1(VTC1) and VTC3 values differed from Time 1 to Time 2 as well as from Time 1 to Time 3. The proportion of VTC1 in samples generally decreased, while VTC3, representative of aperiodicity, increased. Both nonlinear energy difference ratio and spectrum convergence ratio exhibited no significant changes throughout the study. Professional voice training can improve DSI and D-value of vocal range in singers' voices. These parameters have potential to be used for voice training evaluation and screening. Many nonlinear parameters did not detect differences in the healthy voices studied, but VTCPs created using intrinsic dimension present a valuable new method, visualizing increases in aperiodicity of the speaking voices after professional voice training.