Electromagnetic Articulography Research Articles

Temporal Coordination of Articulatory Gestures in Nonnative Onset Clusters: Evidence From American English Speakers Using Electromagnetic Articulography.

Research in cross-language speech production indicates that, although the production of nonnative consonant clusters is often difficult, speakers of American English can produce some nonnative clusters (e.g., /fn/) with high accuracy. This ease of production for select nonnative clusters may occur due to similarity of phonetic structure with native clusters (e.g., nonnative /fn/ and native /sm/ are both fricative-nasal sequences). The current study tested this hypothesis by examining the extent of transfer of articulatory coordination from phonetically similar native onset clusters (i.e., /fl/, /sm/) to nonnative /fn/ clusters. Using electromagnetic articulography, lip, tongue, and jaw movements were recorded in nine participants during the production of 22 nonwords (eight tokens per nonword) containing the native and nonnative clusters in different carrier phrases. We examined the temporal lags between each consonantal gesture in a cluster and the flanking vowel gesture, which were compared to the matched singleton conditions. Analyses revealed that, as in native speech, when the syllable onset became more complex (i.e., CV ➔ CCV [C as consonant, V as vowel]), there was an increase in lag (less temporal overlap) between the leftmost consonantal gesture and the vocalic gesture, whereas there was a decrease in lag (more temporal overlap) between the rightmost consonant and the vocalic gesture (i.e., C-center timing). However, the segmental makeup of the cluster and type of carrier phrase used were also found to influence this change in temporal organization, raising new questions for future research. By and large, the findings are in agreement with the idea that the temporal coordination of articulator movements may be transferred from native clusters to phonetically similar nonnative clusters. However, kinematic measures of a broader range of nonnative clusters in different contexts are needed to fully explore this position.

Decoding the dancing of the tongue: A model-based learning approach to phonetic targets in coarticulationa).

A model synthesizing average frequency components from select sentences in an electromagnetic articulography database has been crafted. This revealed the dual roles of the tongue: its dorsum acts like a carrier wave, and the tip acts as a modulation signal within the articulatory realm. This model illuminates anticipatory coarticulation's subtleties during speech planning. It undergoes rigorous, two-stage optimization: statistical estimation and refinement to depict carryover and anticipation. The model's base, rooted in physiological insights, deciphers carryover targets while its upper layer captures anticipation. Optimization has pinpointed unique phonetic targets for each phoneme, providing deep insights into virtual target formation during speech planning. These simulations, aligning closely with empirical data and marked by a mere 0.18 cm average error, along with extensive listening tests attest to the model's accuracy and enhanced speech synthesis quality.

Speech synthesis from three-axis accelerometer signals using conformer-based deep neural network

Silent speech interfaces (SSIs) have emerged as innovative non-acoustic communication methods, and our previous study demonstrated the significant potential of three-axis accelerometer-based SSIs to identify silently spoken words with high classification accuracy. The developed accelerometer-based SSI with only four accelerometers and a small training dataset outperformed a conventional surface electromyography (sEMG)-based SSI. In this study, motivated by the promising initial results, we investigated the feasibility of synthesizing spoken speech from three-axis accelerometer signals. This exploration aimed to assess the potential of accelerometer-based SSIs for practical silent communication applications. Nineteen healthy individuals participated in our experiments. Five accelerometers were attached to the face to acquire speech-related facial movements while the participants read 270 Korean sentences aloud. For the speech synthesis, we used a convolution-augmented Transformer (Conformer)-based deep neural network model to convert the accelerometer signals into a Mel spectrogram, from which an audio waveform was synthesized using HiFi-GAN. To evaluate the quality of the generated Mel spectrograms, ten-fold cross-validation was performed, and the Mel cepstral distortion (MCD) was chosen as the evaluation metric. As a result, an average MCD of 5.03 ± 0.65 was achieved using four optimized accelerometers based on our previous study. Furthermore, the quality of generated Mel spectrograms was significantly enhanced by adding one more accelerometer attached under the chin, achieving an average MCD of 4.86 ± 0.65 (p < 0.001, Wilcoxon signed-rank test). Although an objective comparison is difficult, these results surpass those obtained using conventional SSIs based on sEMG, electromagnetic articulography, and electropalatography with the fewest sensors and a similar or smaller number of sentences to train the model. Our proposed approach will contribute to the widespread adoption of accelerometer-based SSIs, leveraging the advantages of accelerometers like low power consumption, invulnerability to physiological artifacts, and high portability.

The encoding of prominence relations in supra-laryngeal articulation across speaking styles

Speakers modify their supra-laryngeal articulation on the one hand as a function of prosodic prominence and on the other hand as a function of speaking style. Articulatory movements are spatio-temporally expanded in words that are highlighted due to prosodic strengthening, but also in utterances that are produced in a loud speaking style. What remains unclear, to date, is how these two phenomena interact, or, in other words, how prominence relations are encoded in loud speech. In this paper, we use electromagnetic articulography to investigate lip and tongue kinematics and analyze selected acoustic parameters across several focus conditions in habitual and loud speech. Prosodic prominence relations are studied as the simultaneous strengthening of highlighted words and the attenuation of words in the background. The results show that prominence relations can be encoded multidimensionally in loud speech. Some modifications are even stronger in loud speech than in habitual speech, despite the general articulatory modifications associated with a loud speaking style. In both speaking styles, the attenuation of words in the background tends to be more robust than the strengthening of highlighted words. The paper underlines the flexibility of the multidimensional phonetic space in the encoding of prosodic prominence across varying communicative demands.

Immediate auditory feedback regulates inter-articulator speech coordination in service to phonetic structure.

Research has shown that talkers reliably coordinate the timing of articulator movements across variation in production rate and syllable stress, and that this precision of inter-articulator timing instantiates phonetic structure in the resulting acoustic signal. We here tested the hypothesis that immediate auditory feedback helps regulate that consistent articulatory timing control. Talkers with normal hearing recorded 480 /tV#Cat/ utterances using electromagnetic articulography, with alternative V (/ɑ/-/ɛ/) and C (/t/-/d/), across variation in production rate (fast-normal) and stress (first syllable stressed-unstressed). Utterances were split between two listening conditions: unmasked and masked. To quantify the effect of immediate auditory feedback on the coordination between the jaw and tongue-tip, the timing of tongue-tip raising onset for C, relative to the jaw opening-closing cycle for V, was obtained in each listening condition. Across both listening conditions, any manipulation that shortened the jaw opening-closing cycle reduced the latency of tongue-tip movement onset, relative to the onset of jaw opening. Moreover, tongue-tip latencies were strongly affiliated with utterance type. During auditory masking, however, tongue-tip latencies were less strongly affiliated with utterance type, demonstrating that talkers use afferent auditory signals in real-time to regulate the precision of inter-articulator timing in service to phonetic structure.

Tongue and jaw movement assessed by 3D motion capture during gum chewing.

The tongue plays an important role in mastication, swallowing, and articulation, but it cannot be directly observed because of its location inside the oral cavity. This study aimed to clarify detailed 3D tongue movements during chewing using electromagnetic articulography (EMA). The participants were 10 healthy, young volunteers (average age 26.8 ± 2.1years; 5 males, 5 females). Tongue and jaw movement during gum chewing was measured and recorded using EMA. Four EMA sensors were attached to the anterior, posterior, left, and right surfaces of the tongue, and one sensor was also attached to the mandibular left incisor. The tongue motion during the chewing cycle was spatially and sequentially analyzed based on the motion trajectories of the tongue and mandible. The tongue moved downward and to the masticatory side in a manner similar to the movement of the jaw. The anterior tongue marker moved downward to a greater extent than the other tongue markers. However, the tongue moved forward as the jaw moved backward. The anterior marker reached the most anterior position during the jaw-opening phase and the posterior markers reached the most anterior position during the jaw-closing phase. Just before maximum jaw-opening, all markers on the tongue reached the bottom lowest position. During the jaw-closing phase, the tongue reached the dominant farthest position in the masticatory side. All the markers reached the most posterior position during the occlusal phase. These findings demonstrate the sequence of tongue motion patterns during gum chewing.

Phonotactic and morphonotactic influences on the (a)synchronicity of consonant clusters in Polish

A pilot speech production experiment combined articulatory data obtained using Electromagnetic Articulography (EMA), along with acoustic measures, to investigate the effects of cluster size (CC vs. CCC) and morpheme boundaries on consonant cluster synchronicity for five speakers of Polish. We found that being placed in a larger cluster leads to less synchronous productions of two-consonant sequences. We also found, surprisingly, greater synchronicity for clusters spanning a morpheme boundary than for the same cluster within a morpheme. Our findings may be interpreted from a listener-oriented perspective in which speech production is sensitive to perceptual considerations.

Prosody matters: Preserved prominence marking strategies in people with Parkinson's disease independent of motor status.

While many studies focus on segmental variation in Parkinsonian speech, little is known about prosodic modulations reflecting the ability to adapt to communicative demands in people with Parkinson's disease (PwPD). This type of prosodic modulation is important for social interaction, and it involves modifications in speech melody (intonational level) and articulation of consonants and vowels (segmental level). The present study investigates phonetic cues of prosodic modulations with respect to different focus structures in mild dysarthric PwPD as a function of levodopa. Acoustic and kinematic speech parameters of 25 PwPD were assessed in two motor conditions. Speech production data from PwPD were collected before (medication-OFF) and after levodopa intake (medication-ON) by means of 3-D electromagnetic articulography. On the acoustic level, intensity, pitch, and syllable durations were analyzed. On the kinematic level, movement duration and amplitude were investigated. Spatio-temporal modulations of speech parameters were examined and compared across three different prosodic focus structures (out-of-focus, broad focus, contrastive focus) to display varying speech demands. Overall, levodopa had beneficial effects on motor performance, speech loudness, and pitch modulation. Acoustic syllable durations and kinematic movement durations did not change, revealing no systematic effects of motor status on the temporal domain. In contrast, there were spatial modulations of the oral articulators: tongue tip movements were smaller and lower lip movements were larger in amplitude under levodopa, reflecting a more agile and efficient articulatory movement under levodopa. Thus, respiratory-phonatory functions and consonant production improved, while syllable duration and tongue body kinematics did not change. Interestingly, prominence marking strategies were comparable between the medication conditions under investigation, and in fact, appear to be preserved in mild dysarthric PwPD.

Speech emotion recognition based on bi-directional acoustic–articulatory conversion

Acoustic and articulatory signals are naturally coupled and complementary. The challenge of acquiring articulatory data and the nonlinear ill-posedness of acoustic–articulatory conversions have resulted in previous studies on speech emotion recognition (SER) primarily relying on unidirectional acoustic–articulatory conversions. However, these studies have ignored the potential benefits of bi-directional acoustic–articulatory conversion. Addressing the problem of nonlinear ill-posedness and effectively extracting and utilizing these two modal features in SER remain open research questions. To bridge this gap, this study proposes a Bi-A2CEmo framework that simultaneously addresses the bi-directional acoustic–articulatory conversion for SER. This framework comprises three components: a Bi-MGAN that addresses the nonlinear ill-posedness problem, KCLNet that enhances the emotional attributes of the mapped features, and ResTCN-FDA that fully exploits the emotional attributes of the features. Another challenge is the absence of a parallel acoustic–articulatory emotion database. To overcome this issue, this study utilizes electromagnetic articulography (EMA) to create a multi-modal acoustic–articulatory emotion database for Mandarin Chinese called STEM-E2VA. A comparative analysis is then conducted between the proposed method and state-of-the-art models to evaluate the effectiveness of the framework. Bi-A2CEmo achieves an accuracy of 89.04% in SER, which is an improvement of 5.27% compared with the actual acoustic and articulatory features recorded by the EMA. The results for the STEM-E2VA dataset show that Bi-MGAN achieves a higher accuracy in mapping and inversion than conventional conversion networks. Visualization of the mapped features before and after enhancement reveals that KCLNet reduces the intra-class spacing while increasing the inter-class spacing of the features. ResTCN-FDA demonstrates high recognition accuracy on three publicly available datasets. The experimental results show that the proposed bi-directional acoustic–articulatory conversion framework can significantly improve the SER performance.

Russian assimilatory palatalization is incomplete neutralization

Incomplete neutralization refers to phonetic traces of underlying contrasts in phonologically neutralizing contexts. The present study examines one such context: Russian assimilatory&amp;nbsp;palatalization in C+j sequences. Russian contrasts plain and palatalized consonants, (e.g., /p/ vs. /pj/ with the &amp;ldquo;plain&amp;rdquo; consonants having a secondary articulation, involving retraction of the tongue dorsum (velarization/uvularization)). However, in the C+j sequences (e.g., /pjot/ &amp;lsquo;drink (3ps pres)&amp;rsquo;) the contrast between palatalized and plain consonants is neutralized, due to the palatalization of the plain consonants: /pjot/ &amp;rarr; [pʲjot] (assimilatory palatalization).&amp;nbsp;The purpose of the study is to explore whether the neutralization is complete. To do so, we conducted an electromagnetic articulography (EMA) experiment examining temporal coordination and the spatial position of the tongue body in underlyingly palatalized consonants and those derived from assimilatory palatalization. Articulatory results from four native speakers of Russian revealed that gestures in both conditions are coordinated as complex segments (i.e., they are palatalized consonants); however, there are differences across conditions consistent with the residual presence of a tongue dorsum retraction gesture in the &amp;ldquo;plain&amp;rdquo; obstruents. We conclude that neutralization of the plain-palatal contrast in Russian is incomplete&amp;mdash;consonants in the assimilatory palatalization condition exhibit inter-gestural coordination characteristic of palatalized consonants along with residual evidence of an underlying tongue dorsum retraction (velarization/uvularization) gesture.

Phrase boundaries lacking word prosody: An articulatory investigation of Seoul Korean.

This electromagnetic articulography study explores the kinematic profile of Intonational Phrase boundaries in Seoul Korean. Recent findings suggest that the scope of phrase-final lengthening is conditioned by word- and/or phrase-level prominence. However, evidence comes mainly from head-prominence languages, which conflate positions of word prosody with positions of phrasal prominence. Here, we examine phrase-final lengthening in Seoul Korean, an edge-prominence language with no word prosody, with respect to focus location as an index of phrase-level prominence and Accentual Phrase (AP) length as an index of word demarcation. Results show that phrase-final lengthening extends over the phrase-final syllable. The effect is greater the further away that focus occurs. It also interacts with the domains of AP and prosodic word: lengthening is greater in smaller APs, whereas shortening is observed in the initial gesture of the phrase-final word. Additional analyses of kinematic displacement and peak velocity revealed that Korean phrase-final gestures bear the kinematic profile of IP boundaries concurrently to what is typically considered prominence marking. Based on these results, a gestural coordination account is proposed, in which boundary-related events interact systematically with phrase-level prominence as well as lower prosodic levels, and how this proposal relates to the findings in head-prominence languages is discussed.

Studying components of oral statements of children with dysarthria: a review of Russian and foreign research

The article presents a review of Russian and foreign research devoted to the analysis of various components of oral speech of children with dysarthria. A systematic review of publications from 2000 to 2023 on the problem of diagnosing dysarthria in children using subjective and instrumental methods including a total of 215 studies identified the following analysis criteria: type of article; analyzed components of oral speech (breathing, articulation, phonation, prosody); research methods. The analysis showed that the predominant amount of research devoted to the study of oral speech of children with dysarthria is based on psychological and pedagogical observation, digitalized methods are used less often. Objective speech characteristics of children with dysarthria obtained using instrumental assessment methods are highlighted: kinematic analysis, acoustic spectrographic analysis of speech, electromyography, electromagnetic articulography, electropalatography, etc. Familiarization of speech therapists with modern instrumental methods of examination will expand the understanding of the possibilities of obtaining objective data in the study of speech in children with dysarthria. This will allow us to identify new areas of research involving instrumental methods for the development of modern diagnostic and therapy methodologies to overcome dysarthria in childre

BiLSTM-CNN with fixed weight approach for tracking speech articulatory features

Speech production is a complex sequential process which involve the coordination of various articulatory features. Among them tongue being a highly versatile active articulator responsible for shaping airflow to produce targeted speech sounds that are intellectual, clear, and distinct. This paper presents a novel approach for predicting tongue and lip articulatory features involved in a given speech acoustics using a stacked Bidirectional Long Short-Term Memory (BiLSTM) architecture, combined with a one-dimensional Convolutional Neural Network (CNN) for post-processing with fixed weights initialization. The proposed network is trained with two datasets consisting of simultaneously recorded speech and Electromagnetic Articulography (EMA) datasets, each introducing variations in terms of geographical origin, linguistic characteristics, phonetic diversity, and recording equipment. The performance of the model is assessed in Speaker Dependent (SD), Speaker Independent (SI), corpus dependent (CD) and cross corpus (CC) modes. Experimental results indicate that the proposed model with fixed weights approach outperformed the adaptive weights initialization with in relatively minimal number of training epochs. These findings contribute to the development of robust and efficient models for articulatory feature prediction, paving the way for advancements in speech production research and applications.

Segmental and Syllabic Articulations: A Descriptive Approach.

This article reports on jaw articulation, in connection with the segmental articulations of tongue tip and lips. The jaw is considered a syllable articulator as it opens and closes for each syllable, and the amount of jaw opening is related to syllable strength, given the same phonemic vowel. Here, we investigate the syllabic and the segmental articulations relationships to the acoustic segments. We analyze electromagnetic articulography data of 18 Southern Swedish speakers by testing hypotheses based on the Descriptive Approach to Segmental Articulations. In this approach, articulatory movements are divided into fast intervals and steady states, and the crucial ("primary") articulators' deceleration and acceleration peaks align with the acoustic segment boundaries. Previous findings on the lips and the tongue tip are confirmed. New results are the jaw as noncrucial ("nonprimary") does not align with the acoustic segment boundaries: Its steady states are shorter than consonant and vowel segments. Moreover, place of articulation of the coda affects the jaw cycle, while syllable type only affects jaw closing. No such effects are found on the edges of the jaw cycle, instead only gender affects jaw deceleration at syllable onset. The jaw is not crucial for the timing of the acoustic segment boundaries; also, syllable borders seem to not be affected by place of segmental articulation. The results strengthen the view of the jaw as syllable articulator providing a frame and the other articulators (e.g., lips and tongue tip) as the segmental articulators providing the content.

Generalization of inter-articulator timing control: Evidence from tongue-jaw and lip-jaw kinematics using electromagnetic articulography

Recent research in speech production indicates that talkers reliably control the relative timing of articulator movement onsets across variation in production rate, syllable stress, and segmental makeup and that this precision of inter-articulator timing control instantiates phonetic structure. To date, these timing relations have been highly reliable for tongue-jaw kinematics. In the present study, we address the generality of these timing relations to lip-jaw kinematics. Eleven talkers recorded 240 /tV#Cat/ and 240 / bV#Cab/ utterances using electromagnetic articulography, with alternative V (/ɑ/-/ɛ/) and C (/t/−/d/ or /p/−/b/), across changes in production rate (fast-normal) and stress (first syllable stressed-unstressed). To quantify inter-articulator temporal coordination, the timing of either tongue-tip or lower-lip raising onset for the intervocalic C, relative to the jaw opening-closing cycle for V, was obtained. Results indicate that the same kinematic pattern occurred among both sets of articulators: any manipulation that shortened the jaw opening-closing cycle reduced the latency of either tongue-tip or lower-lip movement onset, relative to the onset of jaw opening. Furthermore, the movement onset latencies of the tongue-tip and lower-lip were both highly differentiated by utterance type, bolstering the view that inter-articulator timing relations instantiate phonetic structure in the resulting acoustic signal.

Effects of motor practice on the temporal coordination of articulatory movements for non-native onset clusters: Kinematic and acoustic evidence using electromagnetic articulography

Research on cross-language speech production has shown that part of the challenge of producing non-native clusters arises from difficulties with temporally coordinating the successive consonantal gestures within a cluster. However, it remains unclear whether the application of practice-based motor learning paradigms can improve or stabilize this aspect of non-native cluster production. This study uses electromagnetic articulography (EMA) to measure the effects of motor practice on the temporal coordination of articulatory movements for non-native onset clusters. Monolingual speakers of American English intensively practiced producing monosyllabic sequences containing non-native onset clusters (e.g., MGAT) over two consecutive days. EMA was used to capture lingual, labial, and jaw motion during successive repetitions. For properly-sequenced cluster repetitions, analysis of the EMA sensor trajectories showed that, over the course of practice, there was a general reduction in inter-gestural timing variability and increased overlap between adjacent consonantal gestures. Furthermore, some of these improvements were maintained to the second day, indicating that subjects started forming durable performance gains. Acoustic analyses also showed a reduction in the duration of epenthetic vowel errors (/mgæt/ → /məgæt/) produced throughout practice. Collectively, the findings suggest that motor practice can improve the temporal coordination of articulatory gestures affiliated with non-native clusters.

Speech perception difficulty modulates theta-band encoding of articulatory synergies.

The human brain tracks available speech acoustics and extrapolates missing information such as the speaker's articulatory patterns. However, the extent to which articulatory reconstruction supports speech perception remains unclear. This study explores the relationship between articulatory reconstruction and task difficulty. Participants listened to sentences and performed a speech-rhyming task. Real kinematic data of the speaker's vocal tract were recorded via electromagnetic articulography (EMA) and aligned to corresponding acoustic outputs. We extracted articulatory synergies from the EMA data with principal component analysis (PCA) and employed partial information decomposition (PID) to separate the electroencephalographic (EEG) encoding of acoustic and articulatory features into unique, redundant, and synergistic atoms of information. We median-split sentences into easy (ES) and hard (HS) based on participants' performance and found that greater task difficulty involved greater encoding of unique articulatory information in the theta band. We conclude that fine-grained articulatory reconstruction plays a complementary role in the encoding of speech acoustics, lending further support to the claim that motor processes support speech perception.NEW & NOTEWORTHY Top-down processes originating from the motor system contribute to speech perception through the reconstruction of the speaker's articulatory movement. This study investigates the role of such articulatory simulation under variable task difficulty. We show that more challenging listening tasks lead to increased encoding of articulatory kinematics in the theta band and suggest that, in such situations, fine-grained articulatory reconstruction complements acoustic encoding.

Control parameters for coordinative structures in speech production

In skilled speech production, sets of articulators work cooperatively to achieve task-specific movement goals, despite rampant contextual variation. Efforts to understand these functional units, termed coordinative structures, have focused on identifying the essential control parameters responsible for allowing articulators to achieve these goals, with some research focusing on temporal parameters (relative timing of movements) and other research focusing on spatiotemporal parameters (phase angle of movement onset for one articulator, relative to another). Here, we compared findings across three recent studies where both types of control parameters were investigated, using electromagnetic articulography recordings. In each study, talkers produced VCV utterances, with alternative V (/ɑ/−/ɛ/) and C (/t/−/d/ or /p/−/ b/), across variation in rate (fast–slow) and stress (first syllable stressed–unstressed). Two measures were obtained: (i) the timing of tongue-tip or lower-lip raising onset for intervocalic C, relative to jaw opening–closing cycles, and (ii) the angle of tongue-tip or lower-lip raising onset, relative to the jaw phase plane. All three studies showed that the correlations of tongue-tip/lower-lip movement onset latencies and jaw opening-closing cycle durations were stronger and more reliable than the correlations of tongue-tip/lower-lip phase angles and jaw opening-closing cycle durations, demonstrating that timing is the critical control parameter.

Determination of Micromovements in Removable Prosthesis during Mastication: A Pilot Study with 3D Electromagnetic Articulography.

Edentulism can generate negative impacts on self-esteem, interpersonal relationships, and oral functions. Removable prostheses are commonly used for tooth replacement, but they may cause discomfort due to micromovements during mastication. Objective and quantifiable methods are needed to evaluate these micromovements. A pilot study was conducted to determine the micromovements in removable prostheses during mastication using a 3D electromagnetic articulography (EMA-3D) system. One elderly participant wearing lower removable prostheses and an upper total well-fitting removable prosthesis was studied. The EMA-3D system was used to record movements in three spatial planes. Peanuts were given as test food, and the participant was instructed to chew normally while recordings were carried out until feeling the need to swallow. Analyses of the upper total prosthesis show micromovements ranging from 0.63 ± 0.11 to 1.02 ± 0.13 mm. During simultaneous analyses of the upper prosthesis and lower partial prosthesis, interference was not observed. This pilot study demonstrated the feasibility of using the EMA-3D system to evaluate micromovements in removable prostheses during mastication. Further research is needed to evaluate a larger sample and assess the clinical implications of these micromovements.

Kinematics of Loud, Soft, and Whispered Speech

Introduction: We examined the effects of four types of voicing on vocal tract movements by comparing the articulatory kinematics of whispered speech to habitual, loud, and soft conditions. Methods: Participants included 10 males and 10 females with no history of communication disorders. They read six stimulus sentences in habitual, loud, soft, and whispered conditions. An electromagnetic articulograph tracked the tongue, jaw, and lip movements. Analysis focused on the words we do from a longer sentence. Vertical tongue/jaw and horizontal lip movements were measured during the production of the retracted and rounded front and back vowels in we do. Results: Soft speech led to smaller and slower lip movements than in the habitual condition. Displacement increased for the tongue and jaw in loud and whispered speech compared to the habitual condition. Tongue and jaw velocity increased for loud but not for whispered speech compared to the habitual condition. Utterance duration increased for loud and whispered conditions. Discussion: The increasing tongue and jaw displacement and velocity from soft to habitual to loud speech reported here is consistent with previous accounts. Whispering was less intense than soft speech, yet it involved larger tongue and jaw movements than habitual speech, possibly reflecting a speaker’s focus on greater articulatory clarity when the acoustic signal is the weakest.