Glottal Features Research Articles

Glottal Features in Vocal Assessment Following Thyroidectomy

ObjectivesVocal dysfunction is a frequent complication following thyroidectomy that can be associated with a negative impact on patient’ quality of life. Although the effect of thyroidectomy on acoustic features has been widely used, the examination of glottal flow characteristics to assess vocal outcomes following thyroid gland surgery has not been included in empirical research, to date. The goal of our study was to evaluate early and short-term vocal outcomes following thyroidectomy based on the analysis of glottal acoustic features during speech production. Study designProspective analytical study Materials and methodsWe evaluated vocal outcomes in patients who underwent thyroidectomy between September 2021 and March 2022. We extracted glottal flow features from their vocal recordings preceding surgery and post-operatively at Day1 and Month1 post-operatively. The extraction of glottal features was performed using a signal processing based approach. We extracted the following features: Open quotients (OQ1 and OQ2), Quasi-open quotient (QoQ), Closing quotient (ClQ), Amplitude quotient (AQ), Normalized Amplitude quotient (NAQ) and Speed quotients (SQ1 and SQ2).We included 39 patients. OQ2 and QoQ decreased significantly at Day1 and Month1. OQ1 and NAQ decreased significantly at Month1. ClQ remained stable at both post-operative assessments. AQ decrease was not significant at both dates. SQ1 increased at Day1 and Month1 but the change was not significant. SQ2 decreased significantly at both Day1 and Month1. OQ, QoQ AQ, NAQ and SQ2 did not recover at Month1. We noted that the decrease of SQ1 and SQ2 correlated significantly with the increase of the Voice Handicap Index-10 (VHI-10) at Month1. ConclusionGlottal acoustic features can be a reliable modality to detect vocal changes following thyroidectomy. Thyroidectomy was associated with a vocal dysfunction that was manifested by the decrease of open, amplitude, and speed quotients. Glottal features can present a potential tool to objectively assess the effect of thyroidectomy on vocal folds movements.

Tense and Lax Vowels in the Lahu Dialect of Yunshan: A Laboratory Phonological Study

A controversy exists regarding whether there are tense-lax vowels in Lahu and whether tense-lax phonation should be accorded phonological status, an issue that is closely related to the physiological movement of the larynx. This study presents an experimental analysis of the acoustic features of the Lahu dialect of Yunshan, using electroglottography to collect acoustic and physiological data and employing VoiceSauce and Eggworks to extract glottal features to explore laryngeal mode of motion. The study demonstrates a distinct tense-lax phonation contrast in Yunshan Lahu, which is characterized by multiple types of phonation, differences between front and back parts of tense vowels, and equal importance of tense-lax contrast and vowel length. To contribute to a systematic description of Tibeto-Burman language, this paper proposes to treat tense-lax contrast as the main feature in phonological induction, using length as its accompanying feature for explanation. From the perspective of the path of historical evolution, tenseness comes from the loss of plosive codas, and is also closely related to voiced, aspirated initials. In the process of transforming tense-lax features to tones, tense vowels retained their features of tenseness as a result of falling tones.

Speech Signals as Biomarkers: using Glottal Features for Non-Invasive COVID-19 Testing

Speech Signals as Biomarkers: using Glottal Features for Non-Invasive COVID-19 Testing

Classification of functional dysphonia using the tunable Q wavelet transform

Functional dysphonia (FD) refers to an abnormality in voice quality in the absence of an identifiable lesion. In this paper, we propose an approach based on the tunable Q wavelet transform (TQWT) to automatically classify two types of FD (hyperfunctional dysphonia and hypofunctional dysphonia) from a healthy voice using the acoustic voice signal. Using TQWT, voice signals were decomposed into sub-bands and the entropy values extracted from the sub-bands were utilized as features for the studied 3-class classification problem. In addition, the Mel-frequency cepstral coefficient (MFCC) and glottal features were extracted from the acoustic voice signal and the estimated glottal source signal, respectively. A convolutional neural network (CNN) classifier was trained separately for the TQWT, MFCC and glottal features. Experiments were conducted using voice signals of 57 healthy speakers and 113 FD patients (72 with hyperfunctional dysphonia and 41 with hypofunctional dysphonia) taken from the VOICED database. These experiments revealed that the TQWT features yielded an absolute improvement of 5.5% and 4.5% compared to the baseline MFCC features and glottal features, respectively. Furthermore, the highest classification accuracy (67.91%) was obtained using the combination of the TQWT and glottal features, which indicates the complementary nature of these features.

Hierarchical Multi-Class Classification of Voice Disorders Using Self-Supervised Models and Glottal Features

Previous studies on the automatic classification of voice disorders have mostly investigated the binary classification task, which aims to distinguish pathological voice from healthy voice. Using multi-class classifiers, however, more fine-grained identification of voice disorders can be achieved, which is more helpful for clinical practitioners. Unfortunately, there is little publicly available training data for many voice disorders, which lowers the classification performance on data from unseen speakers. Earlier studies have shown that the usage of glottal source features can reduce data redundancy in detection of laryngeal voice disorders. Another approach to tackle the problems caused by scarcity of training data is to utilize deep learning models, such as wav2vec 2.0 and HuBERT, that have been pre-trained on larger databases. Since the aforementioned approaches have not been thoroughly studied in the multi-class classification of voice disorders, they will be jointly studied in the present work. In addition, we study a hierarchical classifier, which enables task-wise feature optimization and more efficient utilization of data. In this work, the aforementioned three approaches are compared with traditional mel frequency cepstral coefficient (MFCC) features and one-vs-rest and one-vs-one SVM classifiers. The results in a 3-class classification problem between healthy voice and two laryngeal disorders (hyperfunctional dysphonia and vocal fold paresis) indicate that all the studied methods outperform the baselines. The best performance was achieved by using features from wav2vec 2.0 LARGE together with hierarchical classification. The balanced classification accuracy of the system was 62.77% for male speakers, and 55.36% for female speakers, which outperformed the baseline systems by an absolute improvement of 15.76% and 6.95% for male and female speakers, respectively.

Automatic Assessment of Parkinson's Disease Using Speech Representations of Phonation and Articulation

Speech from people with Parkinson's disease (PD) are likely to be degraded on phonation, articulation, and prosody. Motivated to describe articulation deficits comprehensively, we investigated 1) the universal phonological features that model articulation manner and place, also known as speech attributes, and 2) glottal features capturing phonation characteristics. These were further supplemented by, and compared with, prosodic features using a popular compact feature set and standard MFCC. Temporal characteristics of these features were modeled by convolutional neural networks. Besides the features, we were also interested in the speech tasks for collecting data for automatic PD speech assessment, like sustained vowels, text reading, and spontaneous monologue. For this, we utilized a recently collected Finnish PD corpus (PDSTU) as well as a Spanish database (PC-GITA). The experiments were formulated as regression problems against expert ratings of PD-related symptoms, including ratings of speech intelligibility, voice impairment, overall severity of communication disorder on PDSTU, as well as on the Unified Parkinson's Disease Rating Scale (UPDRS) on PC-GITA. The experimental results show: 1) the speech attribute features can well indicate the severity of pathologies in parkinsonian speech; 2) combining phonation features with articulatory features improves the PD assessment performance, but requires high-quality recordings to be applicable; 3) read speech leads to more accurate automatic ratings than the use of sustained vowels, but not if the amount of speech is limited to correspond to the sustained vowels in duration; and 4) jointly using data from several speech tasks can further improve the automatic PD assessment performance.

Detection of Neurogenic Voice Disorders Using the Fisher Vector Representation of Cepstral Features

Neurogenic voice disorders (NVDs) are caused by damage or malfunction of the central or peripheral nervous system that controls vocal fold movement. In this paper, we investigate the potential of the Fisher vector (FV) encoding in automatic detection of people with NVDs. FVs are used to convert features from frame level (local descriptors) to utterance level (global descriptors). At the frame level, we extract two popular cepstral representations, namely, Mel-frequency cepstral coefficients (MFCCs) and perceptual linear prediction cepstral coefficients (PLPCCs), from acoustic voice signals. In addition, the MFCC features are also extracted from every frame of the glottal source signal computed using a glottal inverse filtering (GIF) technique. The global descriptors derived from the local descriptors are used to train a support vector machine (SVM) classifier. Experiments are conducted using voice signals from 80 healthy speakers and 80 patients with NVDs (40 with spasmodic dysphonia (SD) and 40 with recurrent laryngeal nerve palsy (RLNP)) taken from the Saarbruecken voice disorder (SVD) database. The overall results indicate that the use of the FV encoding leads to better identification of people with NVDs, compared to the defacto temporal encoding. Furthermore, the SVM trained using the combination of FVs derived from the cepstral and glottal features provides the overall best detection performance.

Glottal flow characteristics in vowels produced by speakers with heart failure

Heart failure (HF) is one of the most life-threatening diseases globally. HF is an under-diagnosed condition, and more screening tools are needed to detect it. A few recent studies have suggested that HF also affects the functioning of the speech production mechanism by causing generation of edema in the vocal folds and by impairing the lung function. It has not yet been studied whether these possible effects of HF on the speech production mechanism are large enough to cause acoustically measurable differences to distinguish speech produced in HF from that produced by healthy speakers. Therefore, the goal of the present study was to compare speech production between HF patients and healthy controls by focusing on the excitation signal generated at the level of the vocal folds, the glottal flow. The glottal flow was computed from speech using the quasi-closed phase glottal inverse filtering method and the estimated flow was parameterized with 12 glottal parameters. The sound pressure level (SPL) was measured from speech as an additional parameter. The statistical analyses conducted on the parameters indicated that most of the glottal parameters and SPL were significantly different between the HF patients and healthy controls. The results showed that the HF patients generally produced a more rounded glottal pulse and a lower SPL level compared to the healthy controls, indicating incomplete glottal closure and inappropriate leakage of air through the glottis. The results observed in this preliminary study indicate that glottal features are capable of distinguishing speakers with HF from healthy controls. Therefore, the study suggests that glottal features constitute a potential feature extraction approach which should be taken into account in future large-scale investigations in studying the automatic detection of HF from speech.

Glottal features for classification of phonation type from speech and neck surface accelerometer signals

Glottal source characteristics vary between phonation types due to the tension of laryngeal muscles with the respiratory effort. Previous studies in the classification of phonation type have mainly used speech signals recorded by microphone. Recently, two studies were published in the classification of phonation type using neck surface accelerometer (NSA) signals. However, there are no previous studies comparing the use of the acoustic speech signal vs. the NSA signal as input in classifying phonation type. Therefore, the current study investigates simultaneously recorded speech and NSA signals in the classification of three phonation types (breathy, modal, pressed). The general goal is to understand which of the two signals (speech vs. NSA) is more effective in the classification task. We hypothesize that by using the same feature set for both signals, classification accuracy is higher for the NSA signal, which is more closely related to the physical vibration of the vocal folds and less affected by the vocal tract compared to the acoustical speech signal. Glottal source waveforms were computed using two signal processing methods, quasi-closed phase (QCP) glottal inverse filtering and zero frequency filtering (ZFF), and a group of time-domain and frequency-domain scalar features were computed from the obtained waveforms. In addition, the study investigated the use of mel-frequency cepstral coefficients (MFCCs) derived from the glottal source waveforms computed by QCP and ZFF. Classification experiments with support vector machine classifiers revealed that the NSA signal showed better discrimination of the phonation types compared to the speech signal when the same feature set was used. Furthermore, it was observed that the glottal features showed complementary information with the conventional MFCC features resulting in the best classification accuracy both for the NSA signal (86.9%) and the speech signal (80.6%).

The automatic detection of heart failure using speech signals

Heart failure (HF) is a major global health concern and is increasing in prevalence. It affects the larynx and breathing – thereby the quality of speech. In this article, we propose an approach for the automatic detection of people with HF using the speech signal. The proposed method explores mel-frequency cepstral coefficient (MFCC) features, glottal features, and their combination to distinguish HF from healthy speech. The glottal features were extracted from the voice source signal estimated using glottal inverse filtering. Four machine learning algorithms, namely, support vector machine, Extra Tree, AdaBoost, and feed-forward neural network (FFNN), were trained separately for individual features and their combination. It was observed that the MFCC features yielded higher classification accuracies compared to glottal features. Furthermore, the complementary nature of glottal features was investigated by combining these features with the MFCC features. Our results show that the FFNN classifier trained using a reduced set of glottal + MFCC features achieved the best overall performance in both speaker-dependent and speaker-independent scenarios.

The Detection of Parkinson's Disease From Speech Using Voice Source Information

Developing automatic methods to detect Parkinson's disease (PD) from speech has attracted increasing interest as these techniques can potentially be used in telemonitoring health applications. This article studies the utilization of voice source information in the detection of PD using two classifier architectures: traditional pipeline approach and end-to-end approach. The former consists of feature extraction and classifier stages. In feature extraction, the baseline acoustic features-consisting of articulation, phonation, and prosody features-were computed and voice source information was extracted using glottal features that were estimated by iterative adaptive inverse filtering (IAIF) and quasi-closed phase (QCP) glottal inverse filtering methods. Support vector machine classifiers were developed utilizing the baseline and glottal features extracted from every speech utterance and the corresponding healthy/PD labels. The end-to-end approach uses deep learning models which were trained using both raw speech waveforms and raw voice source waveforms. In the latter, two glottal inverse filtering methods (IAIF and QCP) and zero frequency filtering method were utilized. The deep learning architecture consists of a combination of convolutional layers followed by a multilayer perceptron. Experiments were performed using PC-GITA speech database. From the traditional pipeline systems, the highest classification accuracy (67.93%) was given by combination of baseline and QCP-based glottal features. From the end-to-end-systems, the highest accuracy (68.56%) was given by the system trained using QCP-based glottal flow signals. Even though classification accuracies were modest for all systems, the study is encouraging as the extraction of voice source information was found to be most effective in both approaches.

Automatic assessment of intelligibility in speakers with dysarthria from coded telephone speech using glottal features

In clinical practice, assessment of intelligibility in speakers with dysarthria is performed by speech-language pathologists through auditory perceptual tests which demand patients’ presence at hospital and involve time-consuming examinations. Frequent clinical monitoring can be costly and logistically inconvenient both for patients and medical experts. Here, we aim to automate the procedure of assessment of intelligibility in dysarthric speakers with an objective, speech-based method that can be employed in a telescreening application. The proposed method predicts the level of intelligibility in dysarthric speakers using four levels of speech intelligibility (very low, low, mediocre and high). The study compares several automatic methods to assess the intelligibility level in speakers with dysarthria by utilizing information generated at the level of the vocal folds through glottal features and by using coded telephone speech (i.e. speech that is used in telescreening applications). In addition to the glottal features, the openSMILE features are used as acoustic baseline features. Using features obtained from coded speech utterances and the corresponding intelligibility level labels, multiclass-support vector machine (SVM) classifiers are trained. A separate set of multiclass-SVMs are trained using both individual glottal and acoustic features as well as their combinations. Coded telephone speech is generated with the adaptive multi-rate codec with two operational bandwidths (narrowband and wideband), from utterances of an open database of dysarthric speech (Universal Access-Speech). Experimental results showed good classification accuracies for the glottal features, indicating their effectiveness in the intelligibility level assessment in speakers with dysarthria even in the challenging coded condition. Improvement in classification accuracy was obtained when the glottal features were combined with the openSMILE acoustic features, which validate the complimentary nature of the glottal features.

Glottal toggling in Itunyoso Triqui

Itunyoso Triqui (Oto-Manguean: Mexico) possesses several unique morphological derivations, each of which is typified by a toggling of glottal features at the right edge of the root. Root-final coda /ɦ/ is deleted if it is present on uninflected stems, but inserted if it is absent. This process, traditionally known as a morphophonological exchange rule (c.f. Baerman 2007; de Lacy 2012; Wunderlich 2012), is regular and productive in the language. Moreover, it is the primary exponent of the first person singular, the topical third person, and nominal quantifier morphemes, while tonal alternations are secondary, morpheme-specific exponents. The current paper both provides the first comprehensive description of these patterns in Itun-yoso Triqui and argues two theoretical points. First, Triqui glottal toggling involves a morphophono-logical exchange mapping (/α/ → [β]; /β/ → [α]) which, in coordination with syllable well-formedness conditions, produces a toggling pattern. Second, exchange mappings or rules like the Triqui toggle pose unique problems for parallelist approaches to Optimality Theory but not to serialist approaches which permit intermediate stages of representation, a finding that accords well with the necessity for multiple strata in Triqui word formation.

Detection of Specific Language Impairment in Children Using Glottal Source Features

Developmental dysphasia, also known as specific language impairment (SLI), is a language disorder in children that involves difficulty in speaking and understanding spoken words. Detecting SLI at an early stage is very important for successful speech therapy in children. In this paper, we propose a novel approach based on glottal source features for detecting children with SLI using the speech signal. The proposed method utilizes time- and frequency-domain glottal parameters, which are extracted from the voice source signal obtained using glottal inverse filtering (GIF). In addition, Mel-frequency cepstral coefficient (MFCC) and openSMILE based acoustic features are also extracted from speech utterances. Two machine learning algorithms, namely, support vector machine (SVM) and feed-forward neural network (FFNN), are trained separately for the MFCC, openSMILE and glottal features. A leave-fourteen-speakers-out cross-validation strategy is used for evaluating the classifiers. The experiments are conducted using the SLI speech corpus launched by the LANNA research group. Experimental results show that the glottal parameters contain significant discriminative information required for identifying children with SLI. Furthermore, the complementary nature of glottal parameters is investigated by independently combining these features with the MFCC and openSMILE acoustic features. The overall results indicate that the glottal features when used in combination with MFCC feature set provides the best performance with the FFNN classifier in the speaker-independent scenario.

Glottal Source Information for Pathological Voice Detection

Automatic methods for the detection of pathological voice from healthy speech can be considered as potential clinical tools for medical treatment. This study investigates the effectiveness of glottal source information in the detection of pathological voice by comparing the classical pipeline approach to the end-to-end approach. The traditional pipeline approach consists of a feature extractor and a separate classifier. In the former, two sets of glottal features (computed using the quasi-closed phase glottal inverse filtering method) are used together with the widely used openSMILE features. Using both the glottal and openSMILE features extracted from voice utterances and the corresponding healthy/pathology labels, support vector machine (SVM) classifiers are trained. In building end-to-end systems, both raw speech signals and raw glottal flow waveforms are used to train two deep learning architectures: (1) a combination of convolutional neural network (CNN) and multilayer perceptron (MLP), and (2) a combination of CNN and long short-term memory (LSTM) network. Experiments were carried out using three publicly available databases, including dysarthric (the UA-Speech database and the TORGO database) and dysphonic voices (the UPM database). The performance analysis of the detection system based on the traditional pipeline approach showed best results when the glottal features were combined with the baseline openSMILE features. The results of the end-to-end approach indicated higher accuracies (about 2-3 % improvement in all three databases) when glottal flow was used as the raw time-domain input (87.93 % for UA-Speech, 81.12 % for TORGO and 76.66 % for UPM) compared to using raw speech waveform (85.12 % for UA-Speech, 78.83 % for TORGO and 73.71 % for UPM). The evaluation of both approaches demonstrate that automatic detection of pathological voice from healthy speech benefits from using glottal source information.

Dysarthric speech classification from coded telephone speech using glottal features

This paper proposes a new dysarthric speech classification method from coded telephone speech using glottal features. The proposed method utilizes glottal features, which are efficiently estimated from coded telephone speech using a recently proposed deep neural net-based glottal inverse filtering method. Two sets of glottal features were considered: (1) time- and frequency-domain parameters and (2) parameters based on principal component analysis (PCA). In addition, acoustic features are extracted from coded telephone speech using the openSMILE toolkit. The proposed method utilizes both acoustic and glottal features extracted from coded speech utterances and their corresponding dysarthric/healthy labels to train support vector machine classifiers. Separate classifiers are trained using both individual, and the combination of glottal and acoustic features. The coded telephone speech used in the experiments is generated using the adaptive multi-rate codec, which operates in two transmission bandwidths: narrowband (300 Hz - 3.4 kHz) and wideband (50 Hz - 7 kHz). The experiments were conducted using dysarthric and healthy speech utterances of the TORGO and universal access speech (UA-Speech) databases. Classification accuracy results indicated the effectiveness of glottal features in the identification of dysarthria from coded telephone speech. The results also showed that the glottal features in combination with the openSMILE-based acoustic features resulted in improved classification accuracies, which validate the complementary nature of glottal features. The proposed dysarthric speech classification method can potentially be employed in telemonitoring application for identifying the presence of dysarthria from coded telephone speech.

Characteristics of human auditory model based on compensation of glottal features in speech emotion recognition

The speech signal carries emotional message during its production. With the analysis on relation between sound production and glottis, the paper has introduced the glottal features into speech emotion recognition, proposed the model where the glottis is used for compensation of glottal features, and extracted the feature of Glottal Compensation to Zero Crossings with Maximal Teager Energy Operator (GCZCMT). Two experiments have been designed, including that: firstly, the single emotional speech databases of TYUT and Berlin are respectively used for experiment (the purpose of such experiment is to research the emotion recognition capability of GCZCMT feature, and the experimental results show that GCZCMT feature is a feature possibly and effectively distinguishing emotional state); secondly, this experiment is one of mixing speech database (the purpose of such experiment is to research the emotion recognition capability of GCZCMT feature on ross-database language, and the experimental results show that the database dependency of GCZCMT feature is the minimum, and such feature is more suitable for actual complex language environment, and has the higher practical value.

Assessment of dysarthric speech using Elman back propagation network (recurrent network) for speech recognition

In this paper, the role of speech recognition system in the assessment of dysarthric speech based on a method called Elman back propagation network (EBN) is studied. Dysarthria is a neurological disability that damages the control of motor speech articulators. The persons who suffer from Dysarthria may have speech intelligibility rate which may vary from low (2 %) to high (95 %). EBN is a Recurrent network, here a fully connected neural network is built such that the speech characteristics are represented simultaneously by neuron activation states. It is an efficient self supervised training algorithm. For parametric representation of the speech signal, we used Glottal feature along with mel frequency cepstral coefficients. Then finally the output of both the features is compared after the evaluation process using different neural networks and modeling methods. Evaluation of the proposed method is done on the subset of the Universal Access Research database. The subset consists of 9 dysarthric speakers out of 19 speakers each uttering 100 words repeatedly 3 times. The promising performance of the proposed system can be successfully applied to help the people who work for the voice disorder persons.

Algorithm for Cepstral Analysis and Homomorphic Filtering for Glottal Feature Estimation in Speech Signals

Cepstral analysis and homomorphic filtering, are widely used tools for speech analysis. This paper presents an algorithm for cepstral analysis of speech signals. Homomorphic speech processing and cepstrum estimation is performed using the fast Fourier transform (FFT), its inverse operation (IFFT) and the logarithmic magnitude, thus the estimation of cepstral parameters is reached. Using these characteristics, the calculation of glottal speech features is completed, using homomorphic filtering to separate the source and filter signals from the vocal tract response. The last part of this paper shows the obtained signals in the processing task and the analysis as a result of the algorithm process.

Voice source under cognitive load: Effects and classification

Cognitive load refers to the mental demand experienced while performing a cognitive task. A cognitive load measurement system can potentially be a useful tool for monitoring and enhancing human task performance. In the area of speech-based cognitive load classification, while there are various spectral and vocal tract-based features proposed for classification purposes, there is still a lack of studies that investigate how cognitive load affects the voice source, and whether glottal features are effective in cognitive load classification systems. This work introduces a set of databases that contains both speech and electroglottograph (EGG) data. Using these databases, we present results that provide arguably the first direct insight into how cognitive load affects the voice source. Additionally, we show that glottal-based features carry complementary information with respect to formant-based features, and that fusion between glottal and formant-based systems produces classification results that are comparable with (if not better than) existing baseline systems across three out of five evaluation databases.