Glottal Source Signal Research Articles

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.

Evaluation of Glottal Inverse Filtering Techniques on OPENGLOT Synthetic Male and Female Vowels

Current articulatory-based three-dimensional source–filter models, which allow the production of vowels and diphtongs, still present very limited expressiveness. Glottal inverse filtering (GIF) techniques can become instrumental to identify specific characteristics of both the glottal source signal and the vocal tract transfer function to resemble expressive speech. Several GIF methods have been proposed in the literature; however, their comparison becomes difficult due to the lack of common and exhaustive experimental settings. In this work, first, a two-phase analysis methodology for the comparison of GIF techniques based on a reference dataset is introduced. Next, state-of-the-art GIF techniques based on iterative adaptive inverse filtering (IAIF) and quasi closed phase (QCP) approaches are thoroughly evaluated on OPENGLOT, an open database specifically designed to evaluate GIF, computing well-established GIF error measures after extending male vowels with their female counterparts. The results show that GIF methods obtain better results on male vowels. The QCP-based techniques significantly outperform IAIF-based methods for almost all error metrics and scenarios and are, at the same time, more stable across sex, phonation type, F0, and vowels. The IAIF variants improve the original technique for most error metrics on male vowels, while QCP with spectral tilt compensation achieves a lower spectral tilt error for male vowels than the original QCP.

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.

Investigation and Evaluation of Glottal Flow Waveform for Voice Pathology Detection

Automatic voice pathology detection can provide objective estimation and prevention in the early stages of voice diseases. Glottal flow waveform directly reflects the state of glottal excitation. Extracting acoustic features from glottal source signals may contribute to the detection of pathological voice. To improve the performance of voice pathology detection, this article investigates the contribution of the glottal flow waveform for pathological voice detection by evaluating the classification result using features extracted from raw speech utterances and corresponding glottal flow waveforms. The individual feature sets used are extracted from raw or glottal voice utterances with identical parameter settings, which are openSMILE acoustic features, audio features computed by Moving Picture Experts Group-7 standard and classical glottal source features. In addition, a feature selection method in terms of the wrapper approach is used to combine the single features ranked by using the Fisher discrimination ratio. Voice pathology detection experiments were carried out using Random Forest. The best accuracies of 88.52% for the Saarbrucken Voice database and 100.00% for the Massachusetts Eye and Ear Infirmary database are achieved using the combined feature set extracted from the glottal source signal, with improvement of 0.44-3.13% in the accuracies obtained by using raw speech utterances. Compared to state-of-the-art methods, the proposed method achieves the highest accuracy for the Massachusetts Eye and Ear Infirmary database and an increase of 2.75-17.16% in detection accuracy compared to other conventional pipeline systems for the Saarbrucken Voice databse. The experimental results demonstrate that using glottal flow waveform as source signal can improve the performance of pathological voice detection.

Speech spoofing countermeasures based on source voice analysis and machine learning techniques

Automatic speaker verification (ASV) [7] systems are susceptible to malicious attacks. It discredit the performance of a standard ASV system by increasing its false acceptance rates. This paper presents a new countermeasure for the protection of automatic speaker verification systems from spoofed signals. The new countermeasure is based on the analysis of a sequence of acoustic feature vectors using the glottal inverse filtering. In the proposed method, speech is decomposed into a glottal source signal and model the vocal tract filter through glottal inverse filtering. The IAIF desriptors are constructed and are used as features. Support Vector Machines (SVM) classifier and Extreme learning machine (ELM) are used to classify the obtained features as genuine or spoofed. It is hoped that the proposed method can help to detect the genuine speech from the spoofed one.

Alternating minimisation for glottal inverse filtering

A new method is proposed for solving the glottal inverse filtering (GIF) problem. The goal of GIF is to separate an acoustical speech signal into two parts: the glottal airflow excitation and the vocal tract filter. To recover such information one has to deal with a blind deconvolution problem. This ill-posed inverse problem is solved under a deterministic setting, considering unknowns on both sides of the underlying operator equation. A stable reconstruction is obtained using a double regularization strategy, alternating between fixing either the glottal source signal or the vocal tract filter. This enables not only splitting the nonlinear and nonconvex problem into two linear and convex problems, but also allows the use of the best parameters and constraints to recover each variable at a time. This new technique, called alternating minimization glottal inverse filtering (AM-GIF), is compared with two other approaches: Markov chain Monte Carlo glottal inverse filtering (MCMC-GIF), and iterative adaptive inverse filtering (IAIF), using synthetic speech signals. The recent MCMC-GIF has good reconstruction quality but high computational cost. The state-of-the-art IAIF method is computationally fast but its accuracy deteriorates, particularly for speech signals of high fundamental frequency (F0). The results show the competitive performance of the new method: With high F0, the reconstruction quality is better than that of IAIF and close to MCMC-GIF while reducing the computational complexity by two orders of magnitude.

Assessment of disordered voices based on an optimized glottal source model

In this paper, a method for the assessment of disordered voices is proposed. A feature named mean opening quotient (MOQ) obtained from the glottal source estimation is used as an acoustic cue to summarize the degree of severity of the voice disorder. The analysis method uses the empirical mode decomposition algorithm to estimate the glottal source excitation signal from the speech signal. The logarithm of the magnitude spectrum of the speech signal is decomposed into oscillatory modes, called intrinsic mode functions, that are clustered into two classes, the spectral envelope and the harmonic component. The exploitation of the phase information jointly with the estimated harmonic component enables the estimation of the glottal source signal. An appropriate parametric model is fitted to the estimated glottal source excitation signal. The optimal parameters of the glottal source excitation model from which the MOQ is defined are obtained by using a genetic algorithm. The presented method is tested on a corpus of natural speech including the vowel [a] uttered by 22 normophonic speakers and 229 speakers with different degrees of dysphonia. Experimental results show that the proposed method is very effective for assessing the degree of severity of the voice disorder.

Analysis of glottal source parameters in Parkinsonian speech.

Diagnosis and monitoring of Parkinson's disease has a number of challenges as there is no definitive biomarker despite the broad range of symptoms. Research is ongoing to produce objective measures that can either diagnose Parkinson's or act as an objective decision support tool. Recent research on speech based measures have demonstrated promising results. This study aims to investigate the characteristics of the glottal source signal in Parkinsonian speech. An experiment is conducted in which a selection of glottal parameters are tested for their ability to discriminate between healthy and Parkinsonian speech. Results for each glottal parameter are presented for a database of 50 healthy speakers and a database of 16 speakers with Parkinsonian speech symptoms. Receiver operating characteristic (ROC) curves were employed to analyse the results and the area under the ROC curve (AUC) values were used to quantify the performance of each glottal parameter. The results indicate that glottal parameters can be used to discriminate between healthy and Parkinsonian speech, although results varied for each parameter tested. For the task of separating healthy and Parkinsonian speech, 2 out of the 7 glottal parameters tested produced AUC values of over 0.9.

Voice Source Variation Between Vowels in Male Opera Singers

The theory of nonlinear source-filter interaction predicts that the glottal voice source should be affected by the frequency relationship between formants and partials. An attempt to experimentally verify this theory is presented. Glottal voice source and electrolaryngograph (ELG) signal differences between vowels were analyzed in vowel sequences, sung at four pitches with the same degree of vocal loudness by professional opera singers. In addition, the relationships between such differences and the frequency distance between the first formant (F1) and its closest partial were examined. A digital laryngograph microprocessor was used to simultaneously record audio and ELG signals. The former was inverse filtered, and voice source parameters and formant frequencies were extracted. The amplitude quotient of the derivative of the ELG signal (AQdELG) and the contact quotient were also compared. A one-way repeated-measures ANOVA revealed significant differences between vowels, for contact quotient at four pitches and for maximum flow declination rate (MFDR) at three pitches. For other voice source parameters, differences were found at one or two pitches only. No consistent correlation was found between MFDR and the distance between F1 and its closest partial. The glottal voice source tends to vary between vowels, presumably because of source-filter interaction, but the variation does not seem to be dependent on the frequency distance between F1 and its closest partial.

Glottal Spectral Separation for Speech Synthesis

This paper proposes an analysis method to separate the glottal source and vocal tract components of speech that is called Glottal Spectral Separation (GSS). This method can produce high-quality synthetic speech using an acoustic glottal source model. In the source-filter models commonly used in speech technology applications it is assumed the source is a spectrally flat excitation signal and the vocal tract filter can be represented by the spectral envelope of speech. Although this model can produce high-quality speech, it has limitations for voice transformation because it does not allow control over glottal parameters which are correlated with voice quality. The main problem with using a speech model that better represents the glottal source and the vocal tract filter is that current analysis methods for separating these components are not robust enough to produce the same speech quality as using a model based on the spectral envelope of speech. The proposed GSS method is an attempt to overcome this problem, and consists of the following three steps. Initially, the glottal source signal is estimated from the speech signal. Then, the speech spectrum is divided by the spectral envelope of the glottal source signal in order to remove the glottal source effects from the speech signal. Finally, the vocal tract transfer function is obtained by computing the spectral envelope of the resulting signal. In this work, the glottal source signal is represented using the Liljencrants-Fant model (LF-model). The experiments we present here show that the analysis-synthesis technique based on GSS can produce speech comparable to that of a high-quality vocoder that is based on the spectral envelope representation. However, it also permit control over voice qualities, namely to transform a modal voice into breathy and tense, by modifying the glottal parameters.

HMM-Based Speech Synthesis Utilizing Glottal Inverse Filtering

This paper describes an hidden Markov model (HMM)-based speech synthesizer that utilizes glottal inverse filtering for generating natural sounding synthetic speech. In the proposed method, speech is first decomposed into the glottal source signal and the model of the vocal tract filter through glottal inverse filtering, and thus parametrized into excitation and spectral features. The source and filter features are modeled individually in the framework of HMM and generated in the synthesis stage according to the text input. The glottal excitation is synthesized through interpolating and concatenating natural glottal flow pulses, and the excitation signal is further modified according to the spectrum of the desired voice source characteristics. Speech is synthesized by filtering the reconstructed source signal with the vocal tract filter. Experiments show that the proposed system is capable of generating natural sounding speech, and the quality is clearly better compared to two HMM-based speech synthesis systems based on widely used vocoder techniques.

The time course of laryngeal coarticulation in children: First results

Previous work has suggested that the degree or extent of coarticulation is more extreme in young children than adults. Research in this area has focused primarily on supralaryngeal aspects of coarticulation, using spectral measures such as formant frequencies and fricative centroids. At the same time, studies of adults have found that laryngeal adjustments for voiceless consonants extend well into neighboring vowels, yielding higher values of open quotient and DC flow, and more symmetical pulse shapes, in vowels flanking voiceless as compared to voiced consonants. The current work investigates laryngeal coarticulation in normally developing, English-speaking 4- and 5-year olds. Inverse filtering of the oral airflow is used to approximate the glottal source signal in utterances containing /VpV, VhV, VbV/ sequences. The /VbV/ utterance, which does not require vocal-fold abduction, serves as a control condition to the voiceless consonants. Voice source (open quotient, speed quotient) and aerodynamic (AC and DC flow) quantities are measured over time, and compared between the children and adult females. Along with adding to our understanding of developmental changes in coarticulation, these data will contribute to the literature on differences in voice source properties between children and adults. [Work supported by NIH.]

Spectral characterization of jitter, shimmer, and additive noise in synthetically generated voice signals.

Alteration of the harmonic structure in voice source spectra, taken over at least two periods of the waveform, may occur due to the presence of fundamental frequency (f0) perturbation, amplitude perturbation, additive noise, or changes within the glottal source signal itself. In order to make accurate inferences regarding glottal-flow dynamics or perceptual evaluations based on spectral measurements taken from the acoustic speech waveform, investigation of the spectral features of each aperiodic component is required. Based on a heuristic development involving a consideration of the partial sum of the Fourier series taken for two periods of a jittered, shimmered, and (additive, random) noise-contaminated signal, the corresponding spectral characteristics are hypothesized. Subsequent to this, the Fourier series coefficients are calculated for the two periods in order to test the hypotheses. Definite spectral differences are found for each aperiodic component; based on these findings differential quantitative spectral measurements are suggested. Further supportive evidence is obtained through use of Fourier transform and periodogram-averaged calculations. The analysis is carried out on synthetically generated glottal-pulse waveforms and on radiated speech waveforms. A discussion of the results is given in terms of voice aperiodicity in general and in terms of their implication for future studies involving human voice signals.

Glottal source estimation using a sum of exponentials model

This paper describes an algorithm for simultaneously estimating the parameters of a model for the glottal source and the vocal tract filter. The glottal source signal is described by t.he four‐parameter LF model [Fant et al., Speech Transmis. Lab. Q. Prog. Stat. Rep. (1984)]. The vocal tract is modeled as an all‐pole filter; however, to allow for the effects of source‐tract interaction, a separate model is used in the closed and open glottal phases. Under these assumptions, it is shown that the output speech signal can be modeled as a sum of complex exponential signals. The parameters of the speech signal model are estimated using the algorithms described in Kumaresan and Tufts [IEEE Trans. Acoust. Speech Signal Process. ASSP‐30, 833–840 (1982)]. The electroglottograph signal is used to obtain a preliminary estimate of the location of the closed and open phases in each pitch period. The analysis of steady vowel sounds from several speakers indicates that this method provides a very accurate estimate of th...