Are there facial correlates of Thai syllabic tones?

Journal of Voice, 2012

Phonation onset is important in the maintenance of healthy vocal production for speech and singing. The purpose of this preliminary study was to examine differences in vocal fold vibratory behavior between sung and spoken phonation onset gestures. Given the greater degree of precision required for the abrupt onset sung gestures, we hypothesize that differences exist in the timing and coordination of the vocal fold adductory gesture with the onset of vocal fold vibration. Staccato and German (a modified glottal plosive, so named for its occurrence in German classical singing) onset gestures were compared with breathy, normal, and hard onset gestures, using high-speed digital imaging. Samples were obtained from two subjects with no history of voice disorders (a female trained singer and a male nonsinger). Simultaneous capture of acoustical data confirmed the distinction among gestures. Image data were compared for glottal area configurations, degree of adductory positioning, number of small-amplitude prephonatory oscillations (PPOs), and timing of onset gesture events, the latter marked by maximum vocal fold abduction, maximum adduction, beginning of PPOs, and beginning of steady-state oscillation. Results reveal closer adductory positioning of the vocal folds for the staccato and German gestures. The data also suggest a direct relationship between the degree of adductory positioning and the number of PPOs. Results for the timing of onset gesture events suggest a relationship between discrete adductory positioning and more evenly spaced PPOs. By contrast, the overlapping of prephonatory adductory positioning with vibration onset revealed more unevenly spaced PPOs. This may support an existing hypothesis that less well-defined boundaries interfere with normal modes of vibration of the vocal fold tissue.

The Journal of the Acoustical Society of America, 2013

Movements of the mandible have been shown to correlate with English speech rhythm, and significant differences have been found between native speakers' mandible movements and those of second-language speakers. A simple, inexpensive method of inferring movements of the mandible is to use video tracking of a chin marker during speech. However, since the skin is free to stretch over the mandible, inferences using the chin marker may not always be accurate. This study examines the degree of skin stretching during vowel production in 18 different CVC syllables (using 3 different stop consonants and 2 different vowels) spoken as the middle word in a 3-word utterance. We made electromagnetic articulometer (EMA) recordings of two North American English speakers (1 male, 1 female). The distance was measured between coils placed on the lower incisor and on the skin of the mental protruberance (chin). For both speakers, the distance significantly differed depending on the vowel. The onset C affected the distance for only the female speaker. The coda C did not significantly affect the distance for either speaker. These results need to be taken into account as we continue to develop a method for video recording jaw displacement patterns in running speech.

This study investigated the relationship between clearly produced and plain citation form speech styles and motion of visible articulators. Using state-of-the-art computer-vision and image processing techniques, we examined both front and side view videos of speakers' faces while they recited six English words (keyed, kid, cod, cud, cooed, could) containing various vowels differing in visible articulatory features (e.g., lip spreading, lip rounding, jaw displacement), and extracted measurements corresponding to the lip and jaw movements. We compared these measurements in clear and plain speech produced by 18 native English speakers. Based on statistical analyses, we found significant effects of speech style as well as speaker gender and saliency of visual speech cues. Compared to plain speech, we found in clear speech longer duration, greater vertical lip stretch and jaw displacement across vowels, greater horizontal lip stretch for front unrounded vowels, and greater degree of lip rounding and protrusion for rounded vowels. Additionally, greater plain-to-clear speech modifications were found for male speakers than female speakers. These articulatory movement data demonstrate that speakers modify their speech productions in response to communicative needs in different speech contexts. These results also establish the feasibility of utilizing novel computerized facial detection techniques to measure articulatory movements.

2019 22nd Conference of the Oriental COCOSDA International Committee for the Co-ordination and Standardisation of Speech Databases and Assessment Techniques (O-COCOSDA), 2019

We present a database comprising simultaneous acoustic and articulatory recordings of thirty $\mathrm{V}_{1}\mathrm{CV}_{2}$ nonsense words and forty two Toda words, recorded with an electromagnetic articulograph and spoken by six Toda speakers (two males and four females) at four different speaking rates, namely slow, normal, fast and very fast. The vowels in the $\mathrm{V}_{1}\mathrm{CV}_{2}(\mathrm{V}_{1}\neq \mathrm{V}_{2}$) come from a set of six vowels, namely, /a/, /e/, /i/, /o/, /u/, /y/, where the last vowel is a front rounded vowel in Toda. The consonant in the $\mathrm{V}_{1}\mathrm{CV}_{2}$ stimuli is chosen as /p/ for all the recordings. The articulatory data in the proposed database comprises recording of movements of five articulatory points, namely, upper lip, lower lip, jaw, tongue tip and tongue dorsum in the midsagittal plane. The acoustic and articulatory recordings are made available at 16 kHz and 100 Hz respectively. Boundaries of vowels and consonant in $\mat...

The Journal of the Acoustical Society of America, 2012

Video is a convenient, inexpensive method of recording data for jaw movement during speech. However, when using markers attached to the chin, it is possible that the data will not represent actual mandible motion, because of the skin stretching over the mandibleespecially true for labial consonants. In this study, we made video recordings of L1 and L2 speakers of English saying 5 trials of 34 sentences each, and we automatically measured the distance between paper markers attached to the chin and glasses. We compared jaw opening during syllable nucleus for syllables with and without labial onsets, for L1 and L2 English speakers of various proficiencies. Although speakers must stretch the lower lip upwards for a labial constriction, preliminary results show that there are no statistically significant differences for any speaker's jaw opening during the nucleus of non-labial-versus labial-onset syllables. There is also very little intra-subject variation in the metrical structure (as measured by jaw opening) for a given sentence across trials. However, across-trial variability in the time between jaw movement peaks is a lot less for L1 than for L2, presumably because these L2 speakers have not yet mastered the metrical structure of English.

2014

Interspeech 2019, 2019

According to the segmental anchor hypothesis within the Autosegmental-Metrical approach, tones are aligned with segmental boundaries of consonant and vowels in the acoustic domain. In prenuclear rising pitch accents (LH*), the rise is assumed to occur in the vicinity of the accented syllable it is phonologically associated with. However, there are differences in the alignment patterns within and across languages that cannot be captured within the AM approach. In the present study, we investigate the coordination of tonal and oral constriction gestures within Articulatory Phonology. Therefore, we model the coordination of prenuclear LH* pitch accents in Catalan, Northern and Southern German with respect to syllable production on the basis of recordings with a 2D electromagnetic articulography. We provide an extended coupled oscillators model that allows for balanced and imbalanced coupling strengths. Based on examples, we show that the observed differences in alignment patterns for prenuclear rising pitch accents can be modelled with the same underlying coordinative structures/coupling modes for vocalic and tonal gestures and that surface differences arise from gradient variation in coupling strengths. Index Terms: dynamical systems, tonal alignment, tonal gestures, oral constriction gestures, computational model of variability, imbalanced coupling 1.2. Coordination of tonal and oral gestures in AP Articulatory Phonology [9, 10] decomposes speech into a set of potentially overlapping units, articulatory gestures. The temporal organisation of gestures can be modelled by a

1999

In the framework of Articulatory Phonology, linguistic structures are represented in terms of coordinated articulatory gestures that are organized into a gestural score. Syllable structure can be seen as residing in the phasing of the dynamical gestures. The present study investigates the temporal coordination and cohesion of some articulatory gestures as a function of their respective configuration. It looks at the effect of speaking rate and accent on the articulation of the /kl/ or /skl/ clusters and on the same sequence of consonants separated by a reduced vowel (/k♠l/). Articulatory-phonetic representations inspired by gestural scores were recovered independently for both the acoustic and EPG signals that were recorded for two native American English speakers. Constriction locations and degrees were inferred from the EPG data by dividing the palates into place-of-articulation regions and tabulating the percent of electrodes contacted within these regions. These measurements sup...

2015

We investigated articulatory movements of native Japanese speakers’ productions of non-native consonant clusters, using the WAVE system (NDI Corp.). Four Japanese male speakers pronounced “blat”, “bnat”, “btat”, and “pnat” 10 times each in a carrier sentence. Tongue tip displacement from the first to the second consonant varied greatly among the speakers. An ANOVA with two factors, Speaker and Cluster, was performed, and showed significant main effects of both Speakers (p<0.0001) and Clusters (p<0.0001), and significant interaction (p<0.0001). In addition, comparison with previous data of English speakers indicated that the tongue tip displacement in Japanese speakers was larger than that for English speakers. We suggest that the large displacement from the first to second consonant for Japanese speakers indicates that coarticulation is more difficult for them than it is for English speakers.

2003

This paper describes a method for acquiring data for facial movement analysis and implementation in an animated talking head. We will also show preliminary data on how a number of articulatory and facial parameters for some Swedish vowels vary under the influence of expressiveness in speech and gestures. Primarily we have been concerned in expressive gestures and emotions conveying information that is intended to make the animated agent more "human-like" as described in the objectives of the PF-Star 1 project.