Segmental and Syllabic Articulations: A Descriptive Approach.

This paper reports on a one-to-one aspect of the articulatory-acoustic relationship, explaining how acoustic segment boundaries are a result of the rapid movements of the active articulators. In the acceleration profile, these are identified as acceleration peaks, which can be measured. To test the relationship, consonant and vowel segment durations are compared to articulatory posture intervals based on acceleration peaks, and time lags are measured on the alignment of the segment boundaries to the acceleration peaks. Strong relationships and short time lags are expected when the acceleration peaks belong to crucial articulators, whereas weak relationships are expected when the acceleration peaks belong to non-crucial articulators. The results show that lip posture intervals are indeed strongly correlated with [m], and tongue tip postures are strongly correlated with [n]. This is confirmed by the time lag results, which also reveal that the acoustic boundaries precede the acceleration peaks. Exceptions to the predictions are attributed to the speech material or the joint jaw-lip control unit. Moreover, the vowel segments are strongly correlated with the consonantal articulators while less correlated with the tongue body, suggesting that acceleration of crucial consonantal articulators determines not only consonant segment duration but also vowel segment duration.

CHAPTER OUTLINE In this chapter you will learn about: active and passive articulators; complex articulation; instrumental techniques for investigating place of articulation; the anatomy of the vocal tract; the concept of articulation; the main places of articulation for consonant sounds. KEY TERMS Active articulator Articulation Double articulation Passive articulator Place of articulation Vocal tract Introduction In this chapter we shall look at one of the features used to describe the production of consonant sounds. In Chapter 2 we saw that consonants can differ in voicing. So, for example, [p] is voiceless and [b] is voiced, [s] is voiceless and [z] is voiced, and so on. However, it is fairly easy to see that [p] and [b], for instance, have quite a lot in common. If you say the English words pat and bat while looking in a mirror you will see that the first thing you do for each word is close your lips. It isn't so easy with [s] and [z], but you may be able to feel that for both these sounds the tip of your tongue gets very close to the roof of your mouth not far behind your upper front teeth. Try saying Sue and zoo , making the first consonant very long. The topic of this chapter is where in the vocal tract consonants are produced.

We propose a speech-driven 3D tongue animation system. Firstly, the input speech is analyzed to obtain the phoneme sequence. Next, articulatory movements are predicted from the phoneme sequence using a hidden Markov model (HMM) based framework. The HMMs are trained beforehand using a corpus of human articulatory movements, which are recorded by three electromagnetic articulograph (EMA) sensors glued on the tongue tip, tongue body, and tongue dorsum of a speaker respectively. Finally, the predicted articulatory movements are used to control the deformations of a 3D tongue model. The tongue model is a triangular mesh with three key vertices. The three vertices are chosen so that their positions are in correspondence with the three EMA sensors mentioned above. Our tongue model can achieve various tongue shapes with volume preservation. Experiments show that the generated tongue animations are realistic and synchronize well with the input speech.

Surface electromyograms have been recorded for several consonants with a place of articulation in the alveolar region. Discriminable patterns can be identified for /l/, /s/, and /t/, using four locations on and under the tongue tip. A detailed study was made of these patterns in utterance final position, both singly and in cluster with each other and with a nonlingual consonant /p/. Previous results [J. Acoust. Soc. Am. 34, 743 (A) (1962)] had shown that the pattern for the labiodental consonant /f/ is essentialy the same when it is alone in utterance final position, and when it is clustered with alveolar consonants. The present study permits an assessment of contextual effects on alveolar consonants when they are clustered with a nonalveolar and when they are clustered with each other. [This investigation was supported in part by PHS Research Grant DE 01774 from the National Institute of Dental Research, National Institutes of Health, Public Health Service, U. S. Department of Health, Education, and Welfare.]

The open–close jaw cycle established in running speech often spans the articulation of several segments. Moreover, the extremes of the cycle are associated with opposing segment types, consonants and vowels, which form higher-level linguistic units, syllables, when grouped. The present study sought to determine how the syllable is related to the cycle. A partial relationship may imply that the jaw moves linearly from one segment to another. An isomorphic relationship that jaw movement is planned over the articulation of a group of segments, regardless of segment sequencing within the group. Five speakers repeated one of five syllable types (CV, VC, CVC, CCV, VCC) continuously for several seconds, producing an unbroken sequence of consonants and vowels. Jaw movement and acoustic data were recorded simultaneously and used to determine the temporal boundaries of the cycles and syllables, respectively. Cycle boundaries were defined by the points of maximum closure preceding and following an opening gesture, and syllable boundaries according to changes in frequency and amplitude corresponding to the onset of the initial and final segments in the syllable. A comparison of the relative boundary locations suggested an isomorphic relationship between cycles and syllables that interacted with segment type, syllable shape, and speaker.

The purpose of this experiment was to study both the timing and positional properties of articulatory movements in VCV utterances. Conventional cinefluorographic techniques were used to track the movements of the upper lip, lower lip, jaw, tongue tip, and tongue body of two speakers who read randomized lists of VCV utterances containing the vowels /i,a,u/ and the consonants /p,t,k/, in all possible combinations. Results showed that the timing of articulatory movements in a VCV sequence are constrained by the intervocalic consonant, even if the gesture for the consonant is not a contradictory one. Anticipatory movements toward the second vowel always begin during the closure period of the intervocalic consonant. The appearance of carry over coarticulation effects depends on the phonetic identity of the particular segment or degree of involvement of the articulator. Carry over effects, like anticipatory effects, did not extend beyond an immediately adjacent segment. These findings suggest that the rules governing the segmental input to a speech string might be simpler than present models suggest.

Compensatory articulatory mechanisms preserve intelligibility in prodromal Parkinson's disease

The aim of the present research is to examine the relations between auditory perception and production of specific speech contrasts by children with cochlear implants (CIs) who received their implants before 3 years of age and to examine the hierarchy of abilities for perception and production for consonant and vowel features. The following features were examined: vowel height, vowel place, consonant place of articulation (front and back), continuance, and consonant voicing. Fifteen children (mean age = 4;0 and range 3;2 to 5;11) with a minimum of 18 months of experience with their implants and no additional known disabilities served as participants. Perception of feature contrasts was assessed using a modification of the Online Imitative Speech Pattern Contrast test, which uses imitation to assess speech feature perception. Production was examined by having the children name a series of pictures containing consonant and vowel segments that reflected contrasts of each feature. For five of the six feature contrasts, production accuracy was higher than perception accuracy. There was also a significant and positive correlation between accuracy of production and auditory perception for each consonant feature. This correlation was not found for vowels, owing largely to the overall high perception and production scores attained on the vowel features. The children perceived vowel feature contrasts more accurately than consonant feature contrasts. On average, the children had lower perception scores for Back Place and Continuance feature contrasts than for Anterior Place and Voicing contrasts. For all features, the median production scores were 100%; the majority of the children were able to accurately and consistently produce the feature contrasts. The mean production scores for features reflect greater score variability for consonant feature production than for vowel features. Back Place of articulation for back consonants and Continuance contrasts appeared to be the most difficult features to produce, as reflected in lower mean production scores for these features. The finding of greater production than auditory perception accuracy for five of the six features examined suggests that the children with CIs were able to produce articulatory contrasts that were not readily perceived through audition alone. Factors that are likely to play a role in the greater production accuracy in addition to audition include the lexical and phonetic properties of the words elicited, a child's phonological representation of the words and motor abilities, and learning through oro-tactile, visual, proprioceptive, and kinesthetic perception. The differences among the features examined, and between perception and production, point to the clinical importance of evaluating these abilities in children with CIs. The present findings further point to the utility of picture naming to establish a child's production accuracy, which in turn is necessary if using imitation as a measure of auditory capacity.

Bat populations employ rich vocal repertoires for social communication in addition to emitting sound pulses for echolocation. Acoustic parameters of echolocation pulses can vary with the context in which they are emitted, and also with the individual and across populations as a whole. The acoustic parameters of social vocalizations, or “calls”, also vary with the individual and context, but not much is known about their variation across populations at different geographic locations. Here, we leveraged the detailed acoustic classification of social vocalizations available for the Great Himalayan leaf-nosed bat, Hipposideros armiger, to examine geographic variation in five commonly emitted simple syllable types. We hypothesized that individuals within geographically dispersed populations communicate using spectrographically similar constructs or “syllable types”. We also examined whether call syllables vary discordantly with the correlation pattern observed for echolocation pulses across those same geographic regions. Furthermore, we postulated that the acoustic boundaries of a syllable type are not uniquely constrained to its variation within a particular population of the same subspecies. To test our hypotheses, we obtained recordings of social calls of H. a. armiger from nine locations within the oriental region. These locations were consolidated into five geographic regions based on previously established region-specific differences in the peak frequency of echolocation pulses. A multivariate cluster analysis established that unlike echolocation pulses, syllable types exhibit a relatively large variance. Analysis of this variance showed significant differences in Least Squares Means estimates, establishing significant population-level differences in the multiparametric means of individual syllable types across geographic regions. Multivariate discriminant analysis confirmed the presence of region-specific centroids for different syllable constructs, but also showed a large overlap of their multiparametric boundaries across geographic regions. We propose that despite differences in the population-specific core construct of a syllable type, bats maximize acoustic variation across individuals within a population irrespective of its overflow and overlap with other populations.

Speaking requires the coordinated movements of individual articulators. Understanding each articulator’s contribution to speech is fundamental not only for understanding how speech is produced, but also for optimizing speech assessment and treatment. Our recent work has studied the individual contributions of tongue tip, tongue blade, tongue body front, tongue body back, upper lip, and lower lip movement to speech sound production by tracking the motion of sensors attached on the midline of tongue and lips. An optimal set of articulators (tongue tip, tongue body back, upper lip, and lower lip) has been found. However, the tongue lateral (side)’s contribution to speech is still poorly understood. We therefore investigated the contribution of the tongue lateral region to consonant production by analyzing the motion of a sensor attached to the side of tongue. Repeated productions of 12 consonants (including the lateral approximant /l/) were collected from six native English speakers. Consonant classification accuracy based on articulatory movement data obtained using a support vector machine was used as an indication of contribution level. The results suggest that sagittal movement of the tongue lateral sensor did not significantly benefit consonant classification, over and above the optimal set. Implications of these findings are discussed.

Retroflex sounds are classically defined as produced with the tongue tip curled backward and often in contact behind the alveolar ridge ([1], [2], [3]). The sounds, however, present a great inter-language, inter- and intra-speaker articulatory variation. Retroflex stops in Nepali, an Indo-Aryan language spoken in Nepal, are produced with the tongue tip with no backward curling movement at the alveolar ridge ([4],[5]). For Pokharel [5], this is not a real type of retroflexion, but rather is apico-alveolar with no backward curling of the tongue tip. The aim of this study is to experimentally verify Pokharel’s statement and the originality is to go beyond this claim. We wish to verify whether there is any co-articulation effect while producing the retroflex in different vocalic contexts. We use the direct palatography method to determine the place of articulation. Our articulatory data reveal an important articulatory inter and intra speaker variability. The majority of the retroflex stops realized in our study are sub-apico-post-alveolar in the case of back vowels and apico-alveolar in the case of the front vowels. Index terms : retroflex, retraction, Nepali, apical, alveolar, palatograms, linguograms.

Simultaneous measurements of speech articulator movement and laryngeal activity are desirable for obtaining a picture of the coarticulatory behavior between oral articulators and laryngeal vocal fold behavior. However, since electroglottography (EGG) use the application of a weak current across the larynx, which is in close proximity to the oral vocal tract, there is a possibility that its use may interfere with the sensor voltage fluctuation measurements used in electromagnetic articulography (EMA). In order to investigate any hypothetical interference effect, data was collected using Northern Digital Instruments’ Wave articulograph with Glottal Enterprises’s EG2-PCX electroglottograph. Datasets were collected from one male and one female speaker—both with simultaneous EMA and EGG and, for comparison, with EMA alone. Means and standard deviations of inter-sensor distances for static (i.e., reference) and moving (e.g., tongue tip, lip, and jaw) EMA sensor trajectories were compared across the two collection conditions for possible interference effects. Preliminary results find no discernable differences in inter-sensor distances for either static or moving EMA sensors between the with- and without-EGG conditions. If this finding is maintained for additional speakers, NDI EMA data and GE EGG signals may be collected simultaneously without adverse effects on the measurement. [Work supported by NIH.]

According to recent theoretical accounts of place of articulation perception, global, invariant properties of the stop CV syllable onset spectrum serve as primary, innate cues to place of articulation, whereas contextually variable formant transitions constitute secondary, learned cues. By this view, one might expect that young infants would find the discrimination of place of articulation contrasts signaled by formant transition differences more difficult than those cued by gross spectral differences. Using an operant head-turning paradigm, we found that 6-month-old infants were able to discriminate two-formant stimuli contrasting in place of articulation as well as they did five-formant + burst stimuli. Apparently, neither the global properties of the onset spectrum nor simply the additional acoustic information contained in the five-formant + burst stimuli afford the infant any advantage in the discrimination task. Rather, formant transition information provides a sufficient basis for discriminating place of articulation differences.

This study examines trans-vocalic CVC coarticulation of the coronal (dental, alveolar, retroflex and (alveo-)palatal) stop, nasal and lateral consonantsof Arrernte, an Aboriginal language of Central Australia, using electro-palatographic (EPG) recordings of continuous speech. Coronal consonants are known to be more coarticulatorily resistant than peripheral consonants such as bilabials and velars, and this study teases out the coarticulatory differences between these four coronal places of articulation. Results confirm findings from previous studies that laminal consonants (dental and alveo-palatal) are broadly more coarticulatorily aggressive than apical consonants (alveolars and retroflexes), with the alveolars the least resistant to coarticulation. Dental consonants exert fronting effects on both preceding and following consonants, a result consistent with previous acoustic results for these sounds. Similarly, alveo-palatal consonants show both anticipatory and carryover effects, though their exact coarticulatory effect depends on the affected consonant place and manner. The retroflex consonants exert strong fronting effects on thefollowingconsonant, and it is suggested that this is due to the anterior release of retroflexes following ballistic forward movement during closure. There is also some evidence of a retraction effect of retroflexes onprecedingconsonants. Despite this evidence of coarticulatory aggression, retroflexes are particularly affected by preceding and following palatals: in these cases, the retroflex articulation becomes more forward, and dorsal contact for the retroflex is increased. In sum, by examining speech in an understudied language, we learn that there are still greater complexities in the co-ordination of consonants produced using the tongue tip and blade.

Deterioration of oral function can impair articulation, making it essential to evaluate tongue movement during articulation training. Previous reports have indicated an improvement in motor speech disorders, such as glossectomy-related glide issues, through rapid speech exercises, suggesting the potential benefits of non-invasive training. This study aimed to investigate the effect of tongue training on oral function using tongue twisters in healthy Japanese-speaking women. The participants performed tongue twisters designed to engage both the tongue tip and root, with data collected at baseline, 1 month, and 7 months. Participants were required to complete sentences involving the repeated use of /ta/ and /ka/ sounds. This training was conducted 5 times per week for 1 month. Oral diadochokinesis (ODK) was used to measure the number of repetitions per second of each sound, while ultrasound imaging captured the vertical motion of the tongue during pronunciation. The ODK test was used to evaluate speed and regularity of articulatory movement, and ultrasound measurement to assess vertical motion dynamics of the tongue during pronunciation of /ta/ and /ka/. The results indicated a significant improvement in the number of repetitions per second for both /ta/ and /ka/ after 1 month of training. This increase in repetition frequency was maintained for /ta/ even after 7 months, whereas /ka/ showed no further increase in repetition frequency beyond the 1-month mark. Despite this, vertical distance of tongue movement decreased significantly for both syllables after 1 and 7 months. This suggests that /ta/ benefited from both increased efficiency and speed, whereas /ka/ prioritized refined motor control over frequency. The complexity of posterior tongue elevation in /ka/ may require greater precision, limiting increase in repetition despite improved efficiency. This study demonstrates that regular tongue-twister training effectively refines oral function in young women, with sustained benefits over time. The simplicity and accessibility of this method make it a promising approach to maintaining and enhancing oral and articulatory function.