American English Vowels Research Articles

The role of L1-dialect in L2 production

Abstract This study compared the productions of American English (AE) vowels /i/, /u/, /ɛ/, /ɔ/, /ɪ/, /æ/, /ʊ/, /ʌ/, and /ɑ/ between speakers of two Chinese dialects, the Leshan dialect (LSd) and the Xining dialect (XNd). The participants were native speakers of XNd (n = 120) and LSd (n = 120) with advanced English proficiency. Their vowel productions were compared with those of 22 native AE speakers. The results showed that (1) XNd and LSd participants significantly differed in F1 and/or F2 from AE participants when producing /ɪ/, /æ/, /ʌ/, and /ʊ/ and /ɪ/, /ɑ/, /ɔ/, and /ʌ/, respectively; (2) compared with AE participants, XNd participants had significantly longer durations of /ɪ/, /ʊ/, /ʌ/, /ɛ/, and /æ/, and LSd participants had significantly longer durations of /ɪ/, /ʌ/, and /ɛ/. Overall, the vowels that differed in production from that of native speakers were absent in both the L1 dialects and L1 of participants.

Dynamic acoustic vowel distances within and across dialects.

Vowels vary in their acoustic similarity across regional dialects of American English, such that some vowels are more similar to one another in some dialects than others. Acoustic vowel distance measures typically evaluate vowel similarity at a discrete time point, resulting in distance estimates that may not fully capture vowel similarity in formant trajectory dynamics. In the current study, language and accent distance measures, which evaluate acoustic distances between talkers over time, were applied to the evaluation of vowel category similarity within talkers. These vowel category distances were then compared across dialects, and their utility in capturing predicted patterns of regional dialect variation in American English was examined. Dynamic time warping of mel-frequency cepstral coefficients was used to assess acoustic distance across the frequency spectrum and captured predicted Southern American English vowel similarity. Root-mean-square distance and generalized additive mixed models were used to assess acoustic distance for selected formant trajectories and captured predicted Southern, New England, and Northern American English vowel similarity. Generalized additive mixed models captured the most predicted variation, but, unlike the other measures, do not return a single acoustic distance value. All three measures are potentially useful for understanding variation in vowel category similarity across dialects.

Comparing human and machine's use of coarticulatory vowel nasalization for linguistic classificationa).

Anticipatory coarticulation is a highly informative cue to upcoming linguistic information: listeners can identify that the word is ben and not bed by hearing the vowel alone. The present study compares the relative performances of human listeners and a self-supervised pre-trained speech model (wav2vec 2.0) in the use of nasal coarticulation to classify vowels. Stimuli consisted of nasalized (from CVN words) and non-nasalized (from CVCs) American English vowels produced by 60 humans and generated in 36 TTS voices. wav2vec 2.0 performance is similar to human listener performance, in aggregate. Broken down by vowel type: both wav2vec 2.0 and listeners perform higher for non-nasalized vowels produced naturally by humans. However, wav2vec 2.0 shows higher correct classification performance for nasalized vowels, than for non-nasalized vowels, for TTS voices. Speaker-level patterns reveal that listeners' use of coarticulation is highly variable across talkers. wav2vec 2.0 also shows cross-talker variability in performance. Analyses also reveal differences in the use of multiple acoustic cues in nasalized vowel classifications across listeners and the wav2vec 2.0. Findings have implications for understanding how coarticulatory variation is used in speech perception. Results also can provide insight into how neural systems learn to attend to the unique acoustic features of coarticulation.

Identification training adapted to the learner’s vowel space aids in improving native Japanese speakers’ pronunciation of American English vowels

Native speakers of Japanese learning English as a foreign language often have difficulty pronouncing certain American English (AE) vowels, such as the mid vowels /ae/, /ɑ/, / ^/, /ɔ/, and /ɚ/, which typically are assimilated to native vowel categories. It was shown that vowel identification training which was adapted to the learner’s vowel space could aid in improving the native Japanese speakers’ pronunciation of a set of 12 AE vowels that included the 5 more difficult mid vowels. Estimates of the vocal-tract length (VTL) for 53 learners were based upon the central tendency and range of the four lowest formant frequencies observed in their recordings of five short statements in the Japanese language. Having thus characterized the individual vowel spaces of 53 learners, 33 were placed in a “longer-VTL” group while the remaining 20 were placed in a “shorter-VTL” group. Those in the “longer-VTL” group were given periodic identification training using examples of 12 American English monophthongs in a consonant-vowel-consonant (CVC) context that sounded as if a male AE speaker with similar VTL had uttered them. The “shorter-VTL” learners heard training examples that sounded as if the CVCs were uttered by a female AE speaker with similarly short VTL.

Vowel perception under prominence: Examining the roles of F0, duration, and distributional information.

This study investigates how prosodic prominence mediates the perception of American English vowels, testing the effects of F0 and duration. In Experiment 1, the perception of four vowel continua varying in duration and formants (high: /i-ɪ/, /u-ʊ/, non-high: /ɛ-ae/, /ʌ-ɑ/), was examined under changes in F0-based prominence. Experiment 2 tested if cue usage varies as the distributional informativity of duration as a cue to prominence is manipulated. Both experiments show that duration is a consistent vowel-intrinsic cue. F0-based prominence affected perception of vowels via compensation for peripheralization of prominent vowels in the vowel space. Longer duration and F0-based prominence further enhanced the perception of formant cues. The distributional manipulation in Experiment 2 exerted a minimal impact. Findings suggest that vowel perception is mediated by prominence in a height-dependent manner which reflects patterns in the speech production literature. Further, duration simultaneously serves as an intrinsic cue and serves a prominence-related function in enhancing perception of formant cues.

Vowel-internal cues to vowel quality and prominence in speech perception.

This study examines how variation in F0 and intensity impacts the perception of American English vowels. Both properties vary intrinsically as a function of vowel features in the speech production literature, raising the question of the perceptual impact of each. In addition to considering listeners' interpretation of either cue as an intrinsic property of the vowel, the possible prominence-marking function of each is considered. Two patterns of prominence strengthening in vowels, sonority expansion and hyperarticulation, are tested in light of recent findings that contextual prominence impacts vowel perception in line with these effects (i.e.a prominent vowel is expected by listeners to be realized as if it had undergone prominence strengthening). Across four vowel contrasts with different height and frontness features, listeners categorized phonetic continua with variation in formants, F0 and intensity. Results show that variation in level F0 height is interpreted as an intrinsic cue by listeners. Higher F0 cues a higher vowel, following intrinsic F0 effects in the production literature. In comparison, intensity is interpreted as a prominence-lending cue, for which effect directionality is dependent on vowel height. Higher intensity high vowels undergo perceptual re-calibration in line with (acoustic) hyperarticulation, whereas higher intensity non-high vowels undergo perceptual re-calibration in line with sonority expansion.

Spectral and temporal implementation of Japanese speakers’ English vowel categories: A corpus-based study

This study investigates the predictions of second language (L2) acquisition models — SLM, PAM-L2, and L2LP — on how native (L1) Japanese speakers implement the spectral and temporal aspects of L2 American English vowel categories. Data were obtained from 102 L1 Japanese speakers in the J-AESOP corpus, which also includes nativelikeness judgments by trained phoneticians. Spectrally, speakers judged to be less nativelike showed a strong influence of L1-L2 category assimilation, except L2 /æː/ whose dissimilarity from L1 categories likely triggered new category formation, in accordance with SLM and PAM-L2. However, more nativelike speakers showed vowel spectra similar to those of native English speakers for all vowels, which aligns better with L2LP. Temporally, although speakers tended to assimilate the phonetic length of English vowels to Japanese phonemic length distinctions, segment-level L1-L2 category similarity was not a significant predictor of the speakers’ nativelikeness. Instead, implementation of prosodic-level factors such as stress and phrase-final lengthening were better predictors. The results highlight the importance of suprasegmental factors in successful L2 category learning and also reveal a weakness in current models of L2 acquisition, which focus primarily on the segmental level. The use of speech corpora for L2 research is discussed.

Perceptual Assimilation of American English Vowels by Monolingual and Bilingual Learners in Iran

Perceptual Assimilation of American English Vowels by Monolingual and Bilingual Learners in Iran

The challenges and uncertainties when hand measuring F1/F2 values in child speech: Towards an automated approach

Vowels serve a critical role in understanding the acoustic properties of speech. Since they appear early in speech development, they can be used to assess important speech milestones. Prior literature provides the empirical foundation for understanding expected F1/F2 values for typically developing children. Although the traditional measurement of the first and second formant frequencies is a widely accepted practice, its clinical use is limited because of the challenges and uncertainties that are encountered during the process. Automated tools have been developed to address these limitations, however many of these tools have been reported to have inconsistencies when identifying F1/F2 values for child speech. This study explores the use of three semi-automated measurement methods: Wavesurfer Linear Predictive Coding, Praat formant tracking, and a lab-developed approach. The speech of six, native English-speaking, four-year-old children with no history of hearing or speech disorders was analyzed. 48 words, including eight American English vowels, were measured using the three approaches by two researchers, an undergraduate researcher, and an experienced research mentor. The challenges, procedural observations, and accuracy of the methods are discussed. This project was supported by the SURIEA summer research internship program.

Automated vowel space measurement of young children

Vowel formant frequencies are frequently reported as a measure of the articulatory workspace and have been used to study speech development and speech disorders across the lifespan. It is widely acknowledged that the measurement of F1/F2 frequencies for Vowel Space Area (VSA) estimation is a cumbersome task, wrought with challenges and uncertainties. Although there have been advancements in automated approaches, many of these tools continue to be challenged by child speech due to the anatomical differences and dynamic changes of the supraglottal, glottal, and subglottal regions of the airway during the years of speech development. This study examines automated VSA estimation in 4-year-old children using models that have been adapted to manage the higher fundamental frequencies, wider formant bandwidths, and higher amplitudes of subglottal resonances detected in child speech. The speech of six, native American English-speaking, four-year-old children with no history of hearing or speech disorders were included in this investigation. 48 words, including eight American English vowels, were analyzed using traditional manual methods and the SpeechMark MATLAB toolbox automated VSA estimator. The toolbox automates the generation of a convex hull enclosing the child VSA and computation of the area while accounting for the anatomical/acoustic differences that occur during early speech development.

Vowel duration in the Buckeye corpus

Phonological accounts of American English [W. Labov et al., The Atlas of North American English (Walter de Gruyter, Berlin, 2006)] rely on historical and phonotactic analysis to categorize the vowels as either long (two morae) or short (one mora). The vowels of beat, bait, boot, boat, bought, bite, bout, boy are traditionally identified as long, while bit, bet, bat, book, but, bot are supposedly short. It is a reasonable question whether these vowels are indeed longer and shorter, respectively, in phonetic duration. A study of “General American” [H. A. Rositzke, “Vowel-length in general American speech,” Language 15(2), 99–109 (1939)] examined the durations of the vowels of five subjects in scripted utterances, without controlling for dialect. The findings were mostly in accord with the above categories, with the exceptions that [æ, ɑ] (bat, bot) were both phonetically long. The present study is an effort to replicate Rositzke’s on 341 425 vowel tokens in the Buckeye Corpus of conversational English (40 speakers from Columbus, Ohio). The results essentially confirm that the vowels [æ, ɑ] (bat, bot) are phonetically as long as the traditional “long” vowels. This situation calls the traditional analysis of American English vowel length into question.

The neural timecourse of American English vowel discrimination by Japanese, Russian and Spanish second-language learners of English

AbstractThis study investigated the influence of first language (L1) phoneme features and phonetic salience on discrimination of second language (L2) American English (AE) vowels. On a perceptual task, L2 adult learners of English with Spanish, Japanese or Russian as an L1 showed poorer discrimination of the spectral-only difference between /æ:/ as the oddball (deviant) among frequent /ɑ:/ stimuli compared to AE controls. The Spanish listeners showed a significant difference from the controls for the spectral-temporal contrast between /ɑ:/ and /ʌ/ for both perception and the neural Mismatch Negativity (MMN), but only for deviant /ɑ:/ versus /ʌ/ (duration decrement). For deviant /ʌ/ versus /ɑ:/, and for deviant /æ:/ versus /ʌ/ or /ɑ:/, all participants showed equivalent MMN amplitude. The asymmetrical pattern for /ɑ:/ and /ʌ/ suggested that L2 phonetic detail was maintained only for the deviant. These findings indicated that discrimination was more strongly influenced by L1 phonology than phonetic salience.

Applying adaptive recognition of the learner’s vowel space to English pronunciation training of native speakers of Japanese

When native speakers of Japanese are taught English as a second language, there are difficulties with their training in pronunciation of American English vowels that can be ameliorated though adaptive recognition of the learner’s vowel space. This paper reports on the development of an online Computer-Assisted Language Learning (CALL) environment that provides Japanese learners with customized target utterances of 12 single-syllable words that are synthesized according to an adaptive recognition of the learner’s vowel space. These customized target utterances provide each learner with examples of each of 12 American English monophthongs in consonant-vowel-consonant (CVC) context in order to sound as if they had been uttered by the learners themselves. This adaptive process was incorporated into a successfully developed tool for Computer-Assisted Pronunciation Training (CAPT) which gave more appropriate pronunciation targets to each learner, rather than forcing the learners to attempt to match the formant frequencies of their own utterances to those of the target utterances as produced by a speaker exhibiting a different vowel space (i.e., a speaker with a different vocal tract length).

Cochlear-implant simulation mimics increased age in temporal processing of syllables by young and middle-aged normal-hearing listeners

To better understand the effects of signal quality on phonetic processing, the present study examines the effects of signal degradation on processing of two-syllable sequences by young and middle-aged normal-hearing listeners. Temporal processing of American-English vowel sequences was compared across unprocessed and cochlear-implant (CI) simulation conditions for young listeners and for a CI simulation condition for the middle-aged listeners. Using the method of Fogerty, Humes and Kewley-Port [J. Acoust. Soc. Am. 127, 2509–2520 (2010)], listeners heard 70-ms, resynthesized versions of four syllables (“pit, pet, put, pot”) in a two-syllable temporal-order processing task. Task difficulty was adjusted by increasing or decreasing syllable-onset asynchrony (SOA), i.e., the duration between syllable onsets. SOA thresholds for accuracy of syllable-sequence identification was estimated using the method of constant stimuli on each of four 72-trial blocks. Results for the CI simulation conditions for both listener groups were comparable to those for the next older age-group in Fogerty et al. (2010), i.e., young adult listeners performed similarly to the middle-aged listeners Fogerty et al., and middle-aged listeners performed similarly to the older listeners Fogerty et al. Results will be discussed with regard to implications for phonetic processing of speech in demanding listening conditions and practical implications for CI users.

Effects of Individual difference and L1 phonology on the identification and discrimination of American English vowels by native Japanese and Korean listeners

Native speakers of Japanese and Korean identified and discriminated American English vowels /i, ɪ, ɛ, ae, ɑ, ʌ/ in four different consonantal contexts. These listeners also perceptually assimilated the American English vowels to their respective L1 vowel categories. In another session, these listeners produced /i, ɪ, ɛ, ae, ɑ, ʌ/ in two different conditions. They read aloud words on a list, and they repeated after native speakers' utterances. The results revealed that Korean listeners as well as Japanese listeners rely on durational difference to identify and discriminate /i/ and /ɪ/, but Korean listeners perceptually assimilated /i/ and /ɪ/ as roughly equally good exemplars of Korean high front vowel. Japanese listeners, on the other hand, perceived /ɪ/ as somewhat less ideal exemplar of Japanese high front vowel than /i/. However, those Japanese listeners who were sensitive to spectral differences did not outperform those who were less sensitive. Moreover, Japanese listeners sensitive to the spectral differences often misidentified /ɪ/ for /ɛ/. An MDS analysis revealed that in the identification task, listeners’ L1 phonology exercise a stronger influence than do individual differences, but in discrimination task, individual differences are as strong a factor as listeners’ L1 phonology.

Vowel discrimination in noise with formant enhancement: Effects of hearing loss and aging

The aim of this study was to investigate whether formant enhancement of F2 would improve vowel formant discrimination in noise for three groups of listeners (i.e., older listeners with hearing loss, older listeners with normal hearing, and young listeners with normal hearing). Thresholds of vowel formant discrimination were examined for the F2 frequencies of three American English vowels: /ʌ/, /ɛ/, and /i/ in long-term speech-shaped noise with signal-to-noise ratio of + 6 dB and + 12 dB. F2 frequencies of three vowels were enhanced by 6 dB. Thresholds of vowel formant discrimination were significantly highest for older listeners with hearing loss, while young listeners with normal hearing had lowest thresholds with the older listeners with normal hearing in between. However, discrimination thresholds were improved significantly for all three groups after F2 enhancement was implemented regardless of low or high SNR of background noise. Hearing loss and aging resulted in higher thresholds of vowel formant discrimination in noise and F2 enhancements could improve discrimination performance for all listeners in noise.

The relationship between vowel space area and intelligibility of speech in children with velopharyngeal insufficiency

Speech of individuals with velopharyngeal insufficiency is characterized by nasalized vowels. It is well known that velopharyngeal coupling significantly alters the acoustic spectrum of vowels. Aside from changes in formant amplitude and bandwidth, formant frequencies may be modified. Centralized F1 and lowered F2 due to effects of nasalization could change the dimensions of the quadrilateral space. Such variations might cause the vowel space area to be compressed which might contribute to speech unintelligibility. In pursuing research on speech intelligibility in cleft palate, this study was conducted to address the relation between speech intelligibility and spectral characteristics of American English vowels in nasal speech. The hypotheses is that narrower vowel space will result in higher degrees of unintelligibility. Ten children aged 4-12 years with different degrees of hypernasality repeated words with four corner vowels. F1 and F2 midpoint frequencies were measured, and vowel space areas were calculated using the formula for an irregular quadrilateral. In order to determine intelligibility, seventy naive listeners were asked to orthographically transcribe the sentences produced by the participants. Correlational analysis were used to determine the relationship between the vowel space and intelligibility deficits. Further results will be discussed in the meeting.

Variability of formant values at different time points of vowels

Vowel inherent spectral change research has led to two apparently contradictory conclusions about the appropriate representation of vowels: perception studies support a dual-target model of vowels, while production studies focus on calculating higher-fidelity continuous formant contours. This study explores production data in a new way by analyzing where in vowel duration variability is lowest across tokens. Portions of vowels that are less variable across tokens are assumed reflect greater importance to or planning by speakers. Born-and-raised Wisconsinites aged 18–30 were recorded reading a word-in-frame reading list eliciting 30 tokens of 14 American English vowels in the context /hVd/. F1 and F2 were sampled at 50 evenly spaced intervals for each vowel, covering the entire duration. For every speaker, the standard deviations of F1 and F2 were taken separately at every sample. Variability was analyzed as a function of time point. An upturned parabola pattern suggests a single target, a downturned parabola pattern suggests a dual target, and a horizontal pattern suggests continuous specification. For monophthongs, variability is found to be convex, and it is concave for diphthongs. This analysis does not support a dual-target model for monophthongs.

Case Study of the Acoustic Vowel Space of Iraqi Students English Vowels and Standard American English Vowels

Case Study of the Acoustic Vowel Space of Iraqi Students English Vowels and Standard American English Vowels

Duration as a Phonetic Cue in Native and Non‑Native American English

This vowel study looks at the intricate relationship between spectral characteristics and vowel duration in the context of American English vowels, both from a native speaker (NS) and non-native speaker (NNS) perspective. The non-native speaker cohort is homogeneous in the sense that all speakers have Serbian as their mother tongue, but have been long-time residents of the US. The phonetic context investigated in this study is /bVt/, where V is one of the American English monophthongs /i ɪ u ʊ ε æ ʌ ɔ ɑ/. The results of the acoustic analysis show that the NNS vowels are generally longer than the NS vowels. Furthermore, NNSs neutralise the vowel quality of two tense and lax pairs of vowels, /i ɪ/ and /u ʊ/, and rely more heavily on the phonetic duration when prononuncing them.