Beszédtudomány - Speech Science

Subglottal resonances in older adults

2025-04-15T21:19:41+00:00

The subglottal acoustic input impedance partially consists of a series of resonances that are acoustically excited during voiced speech. Measurement of these subglottal resonances (SGRs) has thus far been restricted almost entirely to young adults and children. However, many aspects of the speech production system are known to change as adults age, and there is a possibility that SGR measurements might have clinical value in older adults. This study examined SGRs measured in 10 adults (5 males, 5 females) aged 50-80 years, and probed the dependence of SGR frequencies on vowel quality, posture, pulmonary function, as well as standing height, sitting height, weight, and age. Previously reported data from young adults were also compared with the new data from older adults. Vowel quality affected the frequency of the first subglottal resonance (Sg1) (Sg1 was higher in [a:] than [i:] or [u:]), and age (older adult vs. younger adult) affected the frequencies of the first and second subglottal resonance (Sg1, Sg2) (both SGRs were lower in older adults). Posture did not affect SGR frequencies, and no other significant relationships were found. The interaction of vowel quality with Sg1 is likely due to acoustic coupling between the subglottal and supraglottal (vocal tract) airways during phonation. The interaction between Sg1 and vowel quality was previously reported to be non-significant in younger adults, and the significant interaction in older adults could be due to age-related changes in laryngeal biomechanics and motor control. Based on previous modeling work, the interaction of age with both Sg1 and Sg2 is most likely due to age-related changes in the geometry and biomechanics of the subglottal airways, but empirical verification of this hypothesis is still needed.

Ultrasound analysis of the lingual gestures in Hungarian speech

2025-04-15T21:20:44+00:00

Tamás Gábor Csapó, as one of the founding members of MTA–ELTE “Lendület” Lingual Articulation Research Group, introduced the technique of ultrasound tongue imaging (UTI) into articulatory research in Hungary. He mastered this technique during a Fulbright Scholarship at Indiana University, Bloomington. He did pioneering work in development of UTI data analysis methodology, silent speech interfaces, and other digital applications of UTI. His achievements are acknowledged and his work is referenced world-wide. In the present paper, after a brief outline of the technique of 2D ultrasound tongue imaging, we review the research and application areas in which Tamás Gábor Csapó made his marks within the fields of articulatory research and articulatory-to-acoustic conversion.

Magnitude of ultrasound probe misalignment measures in controlled conditions - a case study

2025-04-15T21:21:51+00:00

We report a method development study aimed for evaluating the reliability of tongue ultrasound data. We analyse the use mean squared error (MSE) of the means ultrasound recordings as a metric of probe stability. The metric’s performance is evaluated against systematically varied speech materials (fronted articulation versus backed articulations) and probe displacement. The speech materials consist of 54 different /CVCVCV/ utterances in random order produced by one native speaker of Finnish and recorded with a Micro ultrasound setup using Articulate Assistant Advanced. In the fronted condition the vowel is /i/ and consonants are varied among /n,s,t/. In the backed condition the vowel is /o/ and the consonants are varied among /h,k,N/. The probe displacement is both simulated and produced intentionally in the real world. For the latter the 54 utterances were repeated in a second block in a different random order. The differences between the results the two displacement methods indicates that this dual approach merits further study. The results also indicate that varying speech materials may overshadow probe displacement which leads to a tentative recommendation of comparing like with like in speech materials when using MSE to detect probe movement.

Artikulációs beszédszintézis megvalósítása dinamikus ultrahangfelvételek alapján

2025-04-15T21:22:39+00:00

Starting from 2D dynamic ultrasound sources recording the movement of the vocal organs and the speech signal of the speaker in a simultaneous and synchronised manner, we produce machine speech by means of artificial intelligence. As visual objects, we use tongue and palate contours fitted automatically to the anatomic boundaries of the ultrasound images, and for training, we extract geometric information from these contours, as the change of their shape fundamentally describes the movement of the vocal organs during articulation. The geometric data consist of radial distances between the tongue and palate contours and coefficients of the discrete cosine transform of the curves, respectively. Relying on this dataset, parameters connected to the acoustic content of the speech signal are trained by the network. These parameters can be interpreted in the framework of the acoustic tube model of the vocal tract, and according to this, reflection coefficients and areas of the articulation channel are to be trained. In this study, sentences are synthesised using linear predictive coding and the acoustic tube model.

Estimation of second subglottal resonance based on F2 measurements and its application in consonant-vowel classification in Hungarian

2025-04-17T09:31:36+00:00

It has been shown for several languages that subglottal resonances (SGRs) play a dividing role in the
frequency space of consonants and vowels (e.g. vowels are separated into the back-front categories by
the second subglottal resonance). Consonant-vowel transitions are characterized by a regression line
(locus equation), and can be classified into distinct categories in the locus equation space, according
to their place of articulation. Several attempts have shown that the dividing lines between these
categories may be the SGRs. In this paper, the relation between CV transitions in the locus equation
space and the separating role of the subglottal resonances are further investigated. Locus equation
space of one native speaker of Hungarian is examined. Consonant-vowel transitions are classified
based on SGRs estimated from the locus equations of a subset of CV sequences. The hit rates and
false alarm rates of the classification are comparable to a baseline experiment where the subglottal
resonances were measured from accelerometer signal.

Búcsú Csapó Tamás Gábortól

2025-04-17T09:14:42+00:00

XXX

Személyes szösszenet, avagy emlékdarabkák Csapó Tamás Gáborról

2025-04-17T09:33:23+00:00

In memory of Tamás Csapó.

Words are as important as action – In the memory of late Csapó Tamás Gábor

2025-04-17T09:33:58+00:00

Personal thoughts in memoriam Tamás Gábor Csapó.