Target Vowel Research Articles

Target–locus scaling methods for modeling families of formant transitions

An approach is proposed for modeling families of vowel-formant transitions in terms of linear-scaling relationships which tie the transitions into a structured system. The model incorporates the hypotheses that (1) transitions are scaled copies of consonant-specific shapes, and (2) the scale factors are the differences between the vowel targets and consonant loci. The terms “target” and “locus” are adopted for their historical roots in acoustic phonetics, but are used here as relational elements. The model is first evaluated with second-formant transitions from VC syllables (the vowels /ɪ, ε, æ, a, ɔ, u, ʌ, ɝ/ combined with the consonants /b, d, ɡ/), repeated 3 times by one adult-male speaker of American English. The model fits these data to within the level of variation among the speaker's repetitions and, within a comparable level, its linear-scaling structure is validated as an approximate feature of the family of transitions. The model's wider capability is then explored by (1) splitting /ɡ/ into its palatal and velar allophones, (2) representing transition shapes with exponentials, and (3) connecting its parameters and properties to those of locus equations. These results illustrate the model's workability and coherency, and argue for its potential to uncover further systematicity in coarticulation.

Unmasking the acoustic effects of vowel-to-vowel coarticulation: A statistical modeling approach

Coarticulation is a source of acoustic variability for vowels, but how large is this effect relative to other sources of variance? We investigate acoustic effects of anticipatory V-to-V coarticulation relative to variation due to the following C and individual speaker. We examine F1 and F2 from V1 in 48 V1-C♯V2 contexts produced by 10 speakers of American English. ANOVA reveals significant effects of both V2 and C on F1 and F2 measures of V1. The influence of V2 and C on acoustic variability relative to that of speaker and target vowel identity is evaluated using hierarchical linear regression. Speaker and target vowel account for roughly 80% of the total variance in F1 and F2, but when this variance is partialed out C and V2 account for another 18% (F1) and 63% (F2) of the remaining target vowel variability. Multinomial logistic regression (MLR) models are constructed to test the power of target vowel F1 and F2 for predicting C and V2 of the upcoming context. Prediction accuracy is 58% for C-Place, 76% for C-Voicing and 54% for V2, but only when variance due to other sources is factored out. MLR is discussed as a model of the parsing mechanism in speech perception.

On the role of consonants and vowels in visual-word processing: Evidence with a letter search paradigm

Prior research has shown that the search function in the visual letter search task may reflect the regularities of the orthographic structure of a given script. In the present experiment, we examined whether the search function of letter detection was sensitive to consonant-vowel status of a pre-cued letter. Participants had to detect the presence/absence of a previously cued letter target (either vowel or consonant) at the initial, central or final position in a five-letter Spanish word or pseudoword. Results showed a significant effect of consonant-vowel status on letter search function which paralleled the orthographic constraints of Spanish. When searching for a consonant, participants showed faster identification of the initial position compared to the central and last positions. The opposite pattern was found for vowels. This result suggests a differential contribution of consonants and vowels to the identification of the orthographic structure of words, in terms of their relative position in Spanish words.

A new non-linear regression model for formant trajectories in English monosyllables incorporating dual targets for vowels.

A new non-linear regression model is proposed to characterize the formant trajectories of the vocalic portion of English CVC syllables in the data described by Hillenbrand et al. [J. Acoust. Soc. Am. 109, 748–763 (2001)]. The modeling framework builds on work of Broad and Clermont [J. Acoust. Soc. Ame. 81, 155–165 (1987)], wherein formant trajectories were modeled via three additive components: (1) a single vowel target, (2) an exponential approach (in time from onset) toward the vowel target from an initial consonant onset value, and (3) an exponential approach (in time from offset) toward the vowel target from a final consonant offset value. The new model extends this to allow for a dual specification (nucleus + offglide) of the vowel targets [T. Nearey and P. Assmann, J. Acoust. Soc. Am. 80, 1297–1308 (1986)]. Initial results suggests that inclusion of a second vowel target provides substantial reduction (about 40%, 23% and 7% respectively for F1, F2, and F3) of error variance on trajectories averaged across 12 speakers. More detailed statistical analyses of variants of the new model are underway and will be reported for both the data described above and that reported on by Assmann et al. [this meeting].

Comparison of the formant frequencies F3 and F4 on a three-dimensional vowel chart.

We have developed a three-dimensional vowel chart that should fit into a quadrilateral IPA vowel chart better and later we have added more visually effective functions to it. On the chart we can see not only the locations of F1 and F2 with two scales of x and y, where x is F1 and y F2, but also F1, F2, and F3 as well, with three scales of x, y, and z, where x is F1, y F2, and z F3. Thus, this enables us to compare the vowel locations of any language (multilingual vowel presentations), vowels uttered by male with those of female (thus, gender difference), vowels of adults with those of children (thus, age difference), foreign language learners with target vowels, or vowels uttered by the same language learners. This time, we focus more on the roles of the respective formants, particularly on the roles of F3 and F4. It has been clarified that F3 is more related to lip-spreading. We want to show that F4 may be more related to lip-protrusion. For this purpose, we have further added a selection of F3 and F4 functions to the chart. [Work supported by JSPS and Chubu University Grant (A).]

Developing vowel mappings for an interactive voice synthesis system controlled by hand motions.

This study investigates vowel mappings for a voice synthesizer controlled by hand gestures for artistic performance. The vowel targets are on a horizontal plane navigated by the movement of the right hand in front of the performer. Two vowel mappings were explored. In one mapping, the vowels were evenly distributed in a circle to make the vowel targets easier for the performer to find. In the other mapping, the vowels were arranged according to the F2 versus F1 space. Linear hand motions were then made through the vowel space while plotting the formant trajectories. The evenly distributed mapping resulted in formant trajectories that were not monotonic; the F1 and F2 pitch contours varied up and down as the hand carried out the linear motions. This had the unintended result of producing multiple diphthongs. In contrast, the F2 versus F1 mapping enabled the performer to create monotonic formant trajectories and the perception of a single diphthong. The performer found it easier to speak and sing through the system when a single linear hand motion resulted in a single diphthong. [This project was supported by Canada Council for the Arts, Natural Sciences and Engineering Council of Canada, and Media and Graphics Interdisciplinary Centre.]

Korean listeners’ sensitivity to language-specific phonetic details of children and adults’ vowel production in five different languages.

Systematic differences in vowel production across languages have been demonstrated, even for so-called “shared” vowels. Starting from 10 months of age, language-specific vowel production patterns emerge in babbling [Boysson-Bardies et al. (1989)]. Vowel production data from a word repetition task in our laboratory also showed language-specific patterns emerging at 2 years of age. While there is research on adult and infant perception of non-native vowels produced by adults, there is little work on adult perception of children’s non-native vowels of different languages. The current study examined Korean adult listeners’ perception of vowels from five languages, Cantonese, English, Greek, Japanese, and Korean, produced by 2-, 5-year-old children, and adults. Consonant-vowel and vowel only stimuli were extracted from our vowel production data. Target vowels included /i/, /a/, and /u/, following velar and alveolar stops. Twenty native Korean adult listeners were asked to identify the vowels and rated their goodness. Results will be analyzed using both the formant frequencies and listeners’ judgments using a mixed effect model. [Supported by a Fulbright fellowship, NIDCD Grant 02932, and NSF Grant 0729140.]

The acquisition of L3 English vowels by infant German–Italian bilinguals

This paper is concerned with a specific case of L3 acquisition: the starting position for English vowel acquisition by infant German–Italian bilinguals will be investigated in light of prototype theory. The chosen example of triple language contact is characterised by consecutive bilingualism as the basis of L3 acquisition, where the learners' L2 is typologically related to L3. It will be discussed whether extended competence in a Germanic language may constitute the basis for transfer in phonetic acquisition of a further Germanic language. The investigation is embedded into a threefold framework: (1) considering the linguistic plurality of the European Union and its educational and linguistic policy, it examines the status of English as L3 in Europe; (2) according to the learners' psycholinguistic characteristics, it discusses the age issue in L3 phonetic acquisition; and (3) against the background of contrastive analysis, it illustrates differences and similarities between German, Italian and English phonetics, and phonology. Within this framework, infant bilinguals' German and Italian vowels will be phonetically analysed and compared with English target vowels. The implications for L3 phonetic acquisition based on the findings are intended as a preliminary, small-scale investigation in preparation of a wider study on a longitudinal basis. The findings presented here are of interest wherever learners share the characteristics of the study participants: bilinguals that add an L3 to their bilingual competence before completion of the so-called critical period.

An acoustic study of vowels and coarticulation as a function of utterance type: A case of acquired apraxia of speech

This case study examines vowel production and coarticulation patterns in AOS as a function of utterance type. A female speaker with AOS, and a sex-matched control participant repeated target vowels in three conditions: isolated word, statement frames and question frames. The first two formant frequencies (F1 and F2) of the target vowels were measured at their onset and temporal midpoints. The vowel spaces of both speakers were examined using a two-dimensional vowel chart; temporal midpoint formant frequency values of the first formant (F1) were plotted against the difference between F2 and F1 (F2–F1) as a representation of the constricted versus open (F1) and fronted versus backed (F2–F1) quality of vowels. The speaker with AOS produced vowels that deviated to a posterior/backed distribution within the vowel space. Coarticulation patterns were also examined using F2 locus equations. The speaker with AOS exhibited reduced coarticulation patterns in all three conditions, when compared to the control speaker. Furthermore, these effects were amplified with increasing motoric complexity. The data elucidate the nature of the compromised speech control system in a speaker with AOS.

Cross-dialectal variation in formant dynamics of American English vowels

This study aims to characterize the nature of the dynamic spectral change in vowels in three distinct regional varieties of American English spoken in the Western North Carolina, in Central Ohio, and in Southern Wisconsin. The vowels /I, epsilon, e, ae, aI/ were produced by 48 women for a total of 1920 utterances and were contained in words of the structure /bVts/ and /bVdz/ in sentences which elicited nonemphatic and emphatic vowels. Measurements made at the vowel target (i.e., the central 60% of the vowel) produced a set of acoustic parameters which included position and movement in the F1 by F2 space, vowel duration, amount of spectral change [measured as vector length (VL) and trajectory length (TL)], and spectral rate of change. Results revealed expected variation in formant dynamics as a function of phonetic factors (vowel emphasis and consonantal context). However, for each vowel and for each measure employed, dialect was a strong source of variation in vowel-inherent spectral change. In general, the dialect-specific nature and amount of spectral change can be characterized quite effectively by position and movement in the F1 by F2 space, vowel duration, TL (but not VL which underestimates formant movement), and spectral rate of change.

Predicting perceptual outcomes from acoustic measures of vowels in dysarthria: A classification analysis.

Reductions in vowel space presumably obscure the distinctiveness of vowels produced by individuals with dysarthria. This represents a source of intelligibility decrement to the extent that more ambiguous vowels influence access of correct lexical items. The present study sought to examine (1) whether vowel formant frequencies predict perceptual decisions through classification analysis and (2) whether different forms of dysarthria are associated with different classification outcomes. Productions of phrases containing the target vowels /i/, /ɪ/, /ε/, /æ/, /u/, /o/, /ɑ/, and /ʌ/ in strong syllables were obtained from patients whose speech was affected by one of four neurological impairments: Parkinson disease, Huntington disease, amyotrophic lateral sclerosis, or cerebellar degeneration. The first two formant frequencies of each vowel were measured at its midpoint and then subjected to a classification analysis. Classification rules based on the acoustic measurements were used to classify each token as one of the eight target vowels. For each speaker group, the performance patterns obtained by the classification analysis were compared to those made by 15 listeners in an open transcription task. The ability of the classification analysis to predict perceptual outcomes specific to each dysarthric speaker group will be discussed. [Work supported by NIH/NIDCD.]

Perception of complete and incomplete formant transitions in vowels.

In everyday speech, formant transitions rarely reach the canonical frequencies of a target vowel. The perceptual system often compensates for such production undershoots, called vowel reduction (VR), by a perceptual overshoot of the final transition frequencies. The present investigation explored the perceptual parameters and existence region of VR. In a series of experiments a 100-ms steady-state vowel V(1) was followed by a formant transition toward a target vowel V(2). By manipulating both its duration and velocity, in most stimuli the transition was truncated and only seldom reached the target. After being presented with the vowel V(2) before each block of trials, listeners were asked to rate their confidence that the transition actually reached the V(2) target. Transitions along six trajectories connecting the three cardinal vowels /a/, /i/, and /u/ in both directions as well as the transition /ie/ (halfway along the trajectory /ia/) were examined in experiments in which either the duration of the transition was fixed and its velocity was varied or vice-versa. Results confirmed the existence of perceptual overshoot and showed that, at the point a transition short of reaching the vowel V(2) was just perceived as if it had reached the target, transition duration and transition velocity were inversely related. The amount of overshoot was found to be larger for larger V(1)-V(2) distances and shorter trajectory durations. The overshoot could be reliably predicted by a linear model based on three parameters--the extent of V(1)-V(2) distance, transition velocity, and transition acceleration. These findings suggest that the perceptual dynamics of speech relies on mechanisms that estimate the rate of change in the resonant characteristics of the vocal tract.

Producing American English Vowels During Vocal Tract Growth: A Perceptual Categorization Study of Synthesized Vowels

To consider interactions of vocal tract change with growth and perceived output patterns across development, the influence of nonuniform vocal tract growth on the ability to reach acoustic-perceptual targets for English vowels was studied. Thirty-seven American English speakers participated in a perceptual categorization experiment. For the experiment, an articulatory-to-acoustic model was used to synthesize 342 five-formant vowels, covering maximal vowel spaces for speakers at 5 growth stages (from 6 months old to adult). Results indicate that the 3 vowels /i u ae/ can be correctly perceived by adult listeners when produced by speakers with a 6-month-old vocal tract. Articulatory-to-acoustic relationships for these 3 vowels differ across growth stages. For a given perceived vowel category, the infant's tongue position is more fronted than the adult's. Furthermore, nonuniform vocal tract growth influences degree of interarticulator coupling for a given perceived vowel, leading to a reduced correlation between jaw height and tongue body position in infantlike compared with adult vocal tracts. Findings suggest that nonuniform vocal tract growth does not prevent the speaker from producing acoustic-auditory targets related to American English vowels. However, the relationships between articulatory configurations and perceptual targets change from birth to adulthood.

Perceptual normalization for variation in speaking style.

Ladefoged and Broadbent [J. Acoust. Soc. Am. 29, 98–104 (1957)] demonstrated that listeners will shift their categorization of a target vowel as a function of acoustic characteristics of a preceding carrier phrase. These results have been interpreted as an example of perceptual normalization for variability resulting from differences in talker anatomy. The present study examined whether listeners would normalize for acoustic variability resulting from differences in speaking style within a single talker. Two vowel series were synthesized that varied between central and peripheral vowels (the vowels in “beat”‐“bit” and “bod”‐“bud”). Each member of the series was appended to one of four carrier phrases that were spoken in either a “clear” or “reduced” speech style. Participants categorized vowels in these eight contexts. A reliable shift in categorization as a function of speaking style was obtained for three of four phrase sets. This demonstrates that phrase context effects can be obtained with a single talker. However, the directions of the obtained shifts are not reliably predicted on the basis of the speaking style of the talker. Instead, it appears that the effect is determined by an interaction of the average spectrum of the phrase with the target vowel. [Work supported by NIH‐NIDCD.]

Phonetic variability in bilinguals’ acquisition of native-vowel category contrasts.

Previous research has shown that bilingual infants and toddlers follow a slightly different developmental pattern in building and stabilizing some native vowel contrastive categories, compared to monolingual populations. Once in the lexical stage, bilingual toddlers do not always show the expected mispronunciation effect when presented with familiar words in which a vowel change has been introduced. To better understand bilinguals’ protracted process in setting the boundaries for certain, perceptually close, native vowel categories, a phonetic analysis of the input characteristics was undertaken. Speech samples containing target words with Catalan midfront vowel sounds were recorded from two groups of Catalan-speaking mothers from monolingual and bilingual environments, differing in the use of their L2 (Spanish). First and second formant values of the target vowels were obtained and compared between groups. The vowels were always correctly and contrastively produced, but bilinguals’ formant values differed significantly from monolinguals’: categories were closer in space and greater variability in production was found. Young bilinguals’ sensitivity to phonetic variability in the words they hear may constrain the development of neat representations of contrastive vowel categories in the early lexicon. [Work supported by CONSOLIDER 2010 Program (CSD2007-012).]

English vowel perception by native speakers of European Portuguese and Mandarin.

This study investigated the perception of eight American English stressed syllabic monophthongs (/i, I, E, æ, V, O, U, u/) by two groups of nonnative listeners: Portuguese and Chinese. Each group was formed by 15 English as a foreign language (EFL) learners, with an upper‐intermediate level of English proficiency, all undergraduate students at a Portuguese university. An identification test was designed to investigate how the participants would perceive the English vowels, which are not present in their first language (L1) inventory as stressed monophthongs (/I, æ, V, U/ in the case of the Portuguese speakers, and all vowels but /i, u/ in the case of the Mandarin speakers). The large number of target vowels played as stimuli allowed a cross‐language analysis of vowel distribution in the acoustic space: for the two groups of EFL listeners, it was more difficult to perceive the English vowels located in a high density area of the English vowel space than those in less crowded areas. For the two groups of speakers, the vowels /æ/ and /U/ had the greatest misidentification rates.

Identification of synthetic vowels based on time‐varying vocal tract area functions.

Identification accuracy for synthetic vowel stimuli generated with static vocal tract area functions from eight speakers was recently reported (K. Bunton and B. H. Story, JASA, in press). Although vowels were identified with reasonably high accuracy, in many cases neighboring vowels were confused. In the present study, new stimuli were generated based, again, on area functions from the same eight speakers. This time, however, the temporal variation of the vocal tract shape during natural production of isolated vowels was incorporated by developing a formant‐to‐area mapping for each speaker based on principal components analysis. This allowed formant contours extracted from recorded vowels to be mapped onto time‐varying area functions. These were coupled to a voice source model and acoustic waves were propagated with a wave‐reflection vocal tract model to generate vowels. Vowels were identified by listeners using a forced choice paradigm. Results indicated that including variation in vowel tract shapes, and hence formant frequencies, along with natural vowel durations improved identification accuracy relative to the previously tested static versions. It is concluded that these eight sets of area functions generate good renditions of target vowels when changes in vocal tract shape and duration were included. [Research supported by NIH R01‐DC04789.]

Discrimination of four English vowel contrasts by Catalan learners varying in language experience.

PAM‐L2 (Best and Tyler, 2007) hypothesizes that discrimination of L2 sounds can be predicted from the perceptual relatedness of L2 categories to L1 categories. Catalan EFL learners encounter serious difficulties to discriminate some English vowel contrasts, partly because Catalan has a smaller vowel inventory than American English. This study presents data from a perceptual discrimination test in which three groups of EFL learners varying in experience with English were asked to discriminate a series of potentially difficult English vowel pairs. The discrimination task consisted in picking the odd item out of three stimuli which were heard one after the other at 1.3‐s intervals. The target vowel stimuli occurred in CVC syllables produced by six native English talkers. Four vowel contrasts were tested: //opena‐/invv/, //eh‐/aelig/, /i‐/smallI/, and /u‐/smallU/. As predicted by PAM‐L2, “category‐goodness” contrasts /i‐/smallI/ and /u‐/smallU/ were fairly well discriminated, the “single‐category” contrast /o...

Native and non-native clear speech production.

Previous work established that both native and non-native clear speech increase intelligibility for native and proficient non-native listeners. However, non-native clear speech enhanced intelligibility less than native clear speech. In this study, we examine native and non-native conversational and clear speech productions with an eye on the differences in hyperarticulation strategies that may underlie a smaller clear speech intelligibility benefit. Results revealed a slower conversational speaking rate for non-native talkers compared with native talkers. Clear speaking rate was similar for both groups suggesting a limit in the speaking rate decrease in intelligibility enhancing clear speech. The durational distinction between tense and lax vowels was increased less in non-native speech due to the smaller speaking rate decrease in clear speech. Analyses of the stop voice onset time (VOT) showed that non-native talkers lengthened the voiced category in the negative VOT direction, while the voiceless category was lengthened less in clear speech. Finally, examination of vowel space expansion in clear speech revealed differences between the two talker groups reflecting non-native talkers’ inexperience in producing some of the vowel targets. Combined, these results suggest that a complex interaction of articulatory difficulty, proficiency, and native language background determines levels of non-native clear speech intelligibility.

Variability of vowel productions within and between days.

Theories of speech production and speech perception assume that phonetic categories can be characterized by stable properties. For example, the notion of stable vowel targets is used to organize articulation, and such targets could serve as the acoustic basis for recognizing vowel categories. However, there is substantial variation in the acoustic patterns of phonetic categories. The lack of invariance between acoustic properties and the phonetic categories of speech has posed a theoretical problem for understanding human speech perception. Although most theories vary in how acoustic‐phonetic variability is approached, theories treat such variability as inherent in speech production or statistically regular or noise. In order to begin to understand acoustic‐phonetic variability, we examined the naturally occurring variability in speech production over time. Participants’ productions of seven different vowels ([IH], [EE], [EH], [AW], [AH], [AE], [UW]) were recorded in nine sessions: at three specific times...