Suzanne Boyce

Articulatory tradeoffs reduce acoustic variability during American English /r/ production

The American English phoneme /r/ has long been associated with large amounts of articulatory variability during production. This paper investigates the hypothesis that the articulatory variations used by a speaker to produce /r/ in different contexts exhibit systematic tradeoffs, or articulatory trading relations, that act to maintain a relatively stable acoustic signal despite the large variations in vocal tract shape. Acoustic and articulatory recordings were collected from seven speakers producing /r/ in five phonetic contexts. For every speaker, the different articulator configurations used to produce /r/ in the different phonetic contexts showed systematic tradeoffs, as evidenced by significant correlations between the positions of transducers mounted on the tongue. Analysis of acoustic and articulatory variabilities revealed that these tradeoffs act to reduce acoustic variability, thus allowing relatively large contextual variations in vocal tract shape for /r/ without seriously degrading the primary acoustic cue. Furthermore, some subjects appeared to use completely different articulatory gestures to produce /r/ in different phonetic contexts. When viewed in light of current models of speech movement control, these results appear to favor models that utilize an acoustic or auditory target for each phoneme over models that utilize a vocal tract shape target for each phoneme.

Coarticulatory stability in American English /r/.

A number of different researchers have reported a substantial degree of variability in how American English /r/ coarticulates with neighboring segments. Acoustic and articulatory data were used to investigate this variability for speakers of “rhotic” American English dialects. Three issues were addressed: (1) the degree to which the F3 trajectory is affected by segmental context and stress, (2) to what extent the data support a “coproduction” versus a “spreading” model of coarticulation, and (3) the degree to which the major acoustic manifestation of American English /r/—the time course of F3—reflects tongue movement for /r/. The F3 formant trajectory durations were measured by automatic procedure and compared for nonsense words of the form /’waCrav/ and /wa’Crav/, where C indicates a labial, alveolar, or velar consonant. These durations were compared to F3 trajectory durations in /’warav/ and /wa’rav/. In addition, formant values in initial syllables of words with and without /r/ were examined for effect...

A new taxonomy of American English /r/ using MRI and ultrasound

In this work we present preliminary results from a large scale production study of American English liquids. MRI and ultrasound have been used to image 20 subjects producing /r/ and /l/. Subjects were native speakers, representative of the main American dialects, and balanced between sexes. MRI data were collected volumetrically for recovery of three‐dimensional tongue shape. In addition, a short (1 s) midsagittal protocol was used in conjunction with ultrasound scanning to confirm the validity of the sustained production required for volumetric imaging. Our results show tongue shapes not inconsistent with the six main types found by Delattre and Freeman (1968), but their range and variety suggest that either additional canonical shapes or a different organizational principal (such as palatal shape) is motivated. We have also observed consistent patterns of tongue grooving associated with different midsagittal shapes, from the parasagittal data not accessible from the Delattre and Freeman cineradiography....

Declination of fundamental frequency in speakers’ production of parenthetical and main clauses

An experimental study was conducted to investigate fundamental frequency (F0) contours in sentences with and without parenthetical clauses. Sentences consisted of a systematically lengthened parenthetical clause inserted between the subject noun phrase and the verb phrase of the main clause. The middle portion of the parenthetical was lengthened incrementally in each sentence to test the effects of sentence length on fundamental frequency contours of seven test sentences. Computer-aided measurements were made for: F0 peaks of key stressed segments; duration of the main clause, parenthetical clause, and clause-final syllables within each; and pauses immediately preceding and following the parenthetical. Mean results demonstrate: (1) a drop in F0 for the parenthetical clause, well below the main clause declination and forming a separate contour of declination; (2) a sharp rise in F0 on return to the main clause; (3) no effect of longer parenthetical length on final segment durations of either the parenthetical or the main clause; and (4) no effect of increased parenthetical length on main clause duration or pause length. These results suggest that parenthetical clauses are mentally programmed as independent constituents, but are subject to some of the same general declination constraints as main clauses.

Grammatical information effects in auditory word recognition

Abstract Three lexical decision experiments were concerned with the separability of syntactic and semantic processing in spoken word perception. An additional experiment examined the problem of measuring reaction times to a spoken stimulus. Words in the Serbo-Croatian language were used; each stimulus consisted of a noun stem (which was either a meaningful root or a pseudoword stem) plus an inflectional suffix which conveyed information about the nouns grammatical case. Speed of identifying the inflectionally related forms of a noun was a function of differences in their syntactic meanings rather than differences in their physical forms or their actual frequencies of occurrence. In addition, identification of a noun was facilitated when it was preceded by a stimulus carrying predictive inflectional information whether that stimulus was a real adjective or pseudoadjective. The results echo previous findings for word perception in print and provide evidence of essential structural uniformity in the processing of inflection for both spoken and printed words. For both, there is evidence that inflectional processing is modular, at least to the extent that it is independent from semantic processing for the initial portion of its time course.

Acoustic noise characteristics of a 4 Telsa MRI scanner

To quantify the acoustic noise characteristics of a 4 Tesla MRI scanner, and determine the effects of structural acoustics and gradient pulse excitations on the sound field so that feasible noise control measures can be developed.

Acoustic differences between ‘‘bunched’’ and ‘‘retroflex’’ variants of American English /r/

In American English, two articulatory variants of /r/ are commonly found: ‘‘bunched’’ /r/—in which the tongue dorsum is bunched in the region of the palate—and ‘‘retroflex’’—in which the tongue tip is curled such that the underside faces the alveopalate. Use of these variants appears to be idiosyncratic for different speakers. Past studies have failed to find any consistent acoustic difference between the variants [cf. Delattre and Freeman (1965)], but have been hampered by the fact that without articulatory confirmation, it is difficult to know which variant was uttered. In the study reported here, Electromagnetometer data from a single speaker producing /r/ in various phonetic and stress contexts was used to determine which instances were ‘‘bunched’’ and which ‘‘retroflex,’’ and a detailed acoustic study of these productions was carried out. Here, it appears to be the case that ‘‘bunched’’ /r/ occurs only when adjacent to /g/. Preliminary results suggest that for the ‘‘bunched’’ /r/, F4 is close in freq...

The Articulatory Phonetics of /r/ for Residual Speech Errors.

Effective treatment for children with residual speech errors (RSEs) requires in-depth knowledge of articulatory phonetics, but this level of detail may not be provided as part of typical clinical coursework. At a time when new imaging technologies such as ultrasound continue to inform our clinical understanding of speech disorders, incorporating contemporary work in the basic articulatory sciences into clinical training becomes especially important. This is particularly the case for the speech sound most likely to persist among children with RSEs-the North American English rhotic sound, /r/. The goal of this article is to review important information about articulatory phonetics as it affects children with RSE who present with /r/ production difficulties. The data presented are largely drawn from ultrasound and magnetic resonance imaging studies. This information will be placed in a clinical context by comparing productions of typical adult speakers to successful versus misarticulated productions of two children with persistent /r/ difficulties.

Cognitive intervention results in web-based videophone treatment adherence and improved cognitive scores.

Background We report findings from an intervention study using telehealth modalities to determine whether provision of telehealth services can improve access to care and increase adherence to cognitive therapy in veterans with mild traumatic brain injury (TBI) while matching traditional care in terms of outcomes. Material/Methods Veterans who were initially non-adherent to clinic-based cognitive therapy were offered a newly developed treatment. The control participants were selected from patient records of veterans who had completed cognitive treatment and matched to MOPS-VI participants on the basis of age, marital or relationship status, and composite memory index score. Baseline and post-treatment cognitive functioning as assessed by the Test of Memory and Learning 2nd Edition (TOMAL-2) was obtained for all participants. The MOPS-VI modules were designed to increase understanding of TBI and elicit problem-solving skills for attention and memory impairment. Results Sixty-seven percent of veterans (who were assigned to the MOPS-VI treatment group because they were initially non-adherent with the clinic-based treatment) completed the MOPS-VI telemedicine treatment. Results of a two-way analysis of Variance (ANOVA) comparing baseline and follow-up scores on the TOMAL-2 in the MOPS-VI and control groups revealed there was a significant pre-post assessment effect, indicating that participant’s memory and learning improved after treatment for both MOPS-VI and standard treatment groups. There was no significant difference between clinic-based treatment and MOPS-VI therapy. Conclusions Preliminary evidence supports the efficacy of the treatment, defined as increased compliance in completing the treatment program, and improvements in standardized memory and learning test results comparable to those following clinic-based treatment.

Verifying a vocal tract model with a closed side-branch.

In this article an implementation of a vocal tract model and its validation are described. The model uses a transmission line model to calculate pole and zero frequencies for a vocal tract with a closed side-branch such as a sublingual cavity. In the validation study calculated pole and zero frequencies from the model are compared with frequencies estimated using elementary acoustic formulas for a variety of vocal tract configurations.