Similar Vowels Research Articles

A nonsense syllable confusion matrix task in Korean‐English bilingual children.

Compared to extensive studies in bilinguals in adults and very young children, very few studies have investigated speech perception in school‐aged Korean‐English bilingual (KEB) children. Given that English (L2) and Korean (L1) have different contrast systems among consonants and vowels and children have different speech perception processing than adults, school‐aged KEB children are expected to have more complicated patterns of L2 perception. The main goal of the present study was to discern which set of L2 sounds are salient enough not to experience interference from L1. Ten KEB children at age 8–13 years were asked to identify 30 nonsense syllables of L2 sounds. The results showed that dissimilar consonants to Korean (i.e., /f,θ/) are most confusing to perceive than similar consonants with familiar contrasts (/p,k,b,g/) next to similar consonants with unfamiliar contrasts (/,,s,z/). More interestingly, perception performance in vowels showed the opposite pattern: a similar vowel (/ε/) was less correctly identified than unfamiliar vowels (/I,æ/). The present study suggests that L2 consonants, which cannot be mapped to any of the L1 sounds, are the most challenging to perceive. Nonetheless, groups of sounds that are patterned together cannot simply be explained with the L1 influence. For example, /f,θ/ are most confusing to KEB as well as English‐monolingual listeners (Phatak & Allen, 2007); yet confusion of /b/ with /θ/ is unique only to KEBs.

The Phonetic Space of Phonological Categories in Heritage Speakers of Mandarin

In two experiments, we investigated the production of Mandarin and English by heritage speakers of Mandarin in comparison to native Mandarin speakers and late learners. In Experiment 1, speakers in all groups made an F2 distinction between Mandarin and English back vowels, with native Mandarin speakers' vowels in both languages having lower F2 values than those of heritage speakers and late learners. In addition, heritage speakers were found to achieve the greatest separation between similar vowel categories. In Experiment 2, few speakers made a VOT distinction between Mandarin unaspirated and English voiced; however, native Mandarin speakers and heritage speakers did distinguish Mandarin aspirated and English voiceless, both groups putting more distance between the two categories than late learners. Thus, we found that heritage speakers maintain not only language-internal contrasts, but also cross-linguistic contrasts, a result which likely stems from an acute approximation of phonetic norms that occurs during early exposure to both languages.

A vowel identification procedure for gerbils

Mongolian gerbils are small rodents with auditory specializations that provide them with excellent hearing in the human speech range. Our lab has previously developed simple go/no-go response procedures to measure gerbil detection and discrimination of human vowels. Here is presented our new identification procedure that uses a two-alternative, forced-choice, left versus right response. The gerbil is tested in a small wire mesh cage that contains a center platform, a light, and two feeder cups left and right of the light. The gerbil begins a trial by mounting the platform when the light flashes. One of two vowel sounds (e.g. /u/ or /i/) then starts repeating at the rate of 1 per sec. The gerbil is reinforced with food if s/he goes to the left feeder cup for /u/, or to the right feeder cup for /i/. Incorrect responses are punished with timeouts. Gerbils easily learn to differentiate the spectrally-dissimilar vowels /u/ versus /i/, and data will be presented on their ability to differentiate more similar vowels such as /a/ versus /ae/. Gerbil identification ability appears related primarily to second formant frequency differences. A video will be shown of gerbils working on the procedure. [Research supported by NIH.]

Speech intelligibility and the vowel space: Normal speech

Vowel formant frequency values and related vowel space measures have been widely used in the study of speech intelligibility for normal and disordered talkers. The strength of vowel space measures in predicting speech intelligibility for normal talkers, however, is relatively low: [Bradlow et al. (1996)] found that vowel space dispersion predicted 19% of variance in sentence intelligibility, and [Hazan and Markham (2004)] found that the F2 difference between /i/ and /u/ predicted 16% of variance in word intelligibility scores. Using data from the 45 men and 48 women speakers in the Hillenbrand et al. (1995) database, acoustic characteristics of ten vowels were used to predict identification accuracy. Multiple regressions revealed that global measures (mean f0, F1 and F2, duration, and amount of formant movement) and fine-grained measures (vowel space area, mean distance among vowels, ranges for f0, F1, and F2, duration ratio between long and short vowels, and dynamic ratio between dynamic and static vowels) accounted for less than one-quarter of the variance in identification scores across talkers. Focusing on confusions among spectrally similar vowels may provide better information on intelligibility differences among normal talkers. Goodness ratings may provide a wider range of scores for statistical analysis than identification accuracy.

Plasticity in vowel perception and production: A study of accent change in young adults

This study investigated changes in vowel production and perception among university students from the north of England, as individuals adapt their accent from regional to educated norms. Subjects were tested in their production and perception at regular intervals over a period of 2 years: before beginning university, 3 months later, and at the end of their first and second years at university. At each testing session, subjects were recorded reading a set of experimental words and a short passage. Subjects also completed two perceptual tasks; they chose best exemplar locations for vowels embedded in either northern or southern English accented carrier sentences and identified words in noise spoken with either a northern or southern English accent. The results demonstrated that subjects at a late stage in their language development, early adulthood, changed their spoken accent after attending university. There were no reliable changes in perception over time, but there was evidence for a between-subjects link between production and perception; subjects chose similar vowels to the ones they produced, and subjects who had a more southern English accent were better at identifying southern English speech in noise.

Vowel identification and vowel space characteristics

Acoustic characteristics of ten vowels produced by 45 men and 48 women from the Hillenbrand et al. (1995) study were correlated with identification accuracy. Global (mean f0, F1 and F2, duration, and amount of formant movement) and distinctive measures (vowel space area, mean distance among vowels, f0, F1 and F2 ranges, duration ratio between long and short vowels, and dynamic ratio between dynamic and static vowels) were used to predict identification scores. Global and distinctive measures accounted for less than one-fourth of variance in identification scores: vowel space area alone accounted for 9% to 12% of variance. Differences in vowel identification across talkers were largely due to poor identification of two spectrally similar vowel pairs /ae/-/eh/ and /uh/-/ah/. Results of acoustic analysis and goodness ratings for well-identified and poorly identified versions of these vowels will be presented. Preliminary analysis revealed that well-identified vowels differed from poorly identified tokens not only in static formant frequencies but also in duration and amount of formant movement over time. A talker’s ability to create distinctions among neighboring vowels is more important in determining vowel intelligibility than overall measures such as vowel space area.

My voice looks like that? Analysis and synthesis of vowel sounds in a PC-based lab

How is it that we can recognize familiar vowel sounds regardless of who is speaking? How does my voice compare to your voice? The acoustics laboratory exercise described here appeals to our natural curiosity about everyday phenomena and about ourselves. In the first part, students use a PC equipped with a built‐in microphone and easy‐to‐use sound analysis software to record themselves ‘‘singing’’ a variety of vowel sounds, each with the same steady pitch. With the spectrum of each vowel displayed, students initially confirm the presence of the harmonic series and identify the fundamental frequency, reinforcing the association of these characteristics with the pitch of the vowel sound. Working in small groups, students then investigate a correlation between the specific vowel and its spectrum envelope. By comparing the spectra of similar vowels sung by different students, as well as the spectra of different vowels, students learn to recognize formants within the spectrum. In the second part, students apply their newfound understanding of the role of formants by attempting to synthesize a vowel sound by adding harmonic components. The sound synthesis is also done with commonly available PC software.

Comparing vowel formant thresholds from two tasks: Classification versus 2‐alternative forced choice (2AFC) adaptive tracking

Accurate classification of vowels in American English is difficult because of the number of acoustically similar vowels. Previous experiments in our lab [Kewley‐Port et al., J. Acoust. Soc. Am. 118, 1929–1930 (2005)] described a new method to obtain formant thresholds using a classification task rather than the more typical 2AFC task. The goal of the present experiment was to directly compare formant thresholds estimated from a classification task with a modified 2AFC paradigm using adaptive tracking. Stimuli, generated by STRAIGHT, consisted of separate continua for formants F1 and F2 based on natural productions of ‘‘bid’’ shifted to ‘‘bed,’’ and ‘‘cut’’ shifted to ‘‘cot.’’ Separate 7‐ and 14‐step continua were created for the classification and 2AFC tasks, respectively. Eight subjects participated first in the classification task, followed by the 2AFC task. Classification threshold results here replicated those observed in our previous experiments. Within‐subject comparisons of the classification and 2AFC data refined our method for calculating a classification threshold from logistic psychometric functions. Individual thresholds calculated using this refined method were very similar to those from the 2AFC task. Results demonstrate the viability of using a more natural, single‐interval classification task to estimate formant discrimination thresholds. [Work supported by Grant NIHDCD‐02229.]

CORRECTNESS, PLEASANTNESS, AND DEGREE OF DIFFERENCE RATINGS ACROSS REGIONS

This study replicates Dennis Preston's folk-dialectology research to determine how speakers from Memphis, Tennessee, view their speech and that of speakers from Northern and Western states on correctness and pleasantness dimensions and how regional and ethnic speech stereotypes affect their evaluations. Results suggest that Memphians consider each region significantly different from the others on language correctness scales and that Southern regions are less “correct” than other regions. However, no regions are rated significantly different for language pleasantness. Such a finding, which elevates regional pleasantness despite regional incorrectness, is not surprising given similar findings elsewhere but does contradict Memphians' behavior when listening to actual speech samples of Southern-shifted vowel tokens where they rate their own speech least educated and pleasant. Intra-Southern differences in ratings on these folk-dialectology tasks, those that appear to demark rural-versus-urban association, are explored as potentially underlying this contrast in behavior. In addition, while they show productively similar vowel systems in Memphis, differences in dialect perceptions between Southern African American and white raters are explored. Finally, to see how Western dialect experience affects participants' pleasantness and correctness ratings, the article also examines how state ratings differed for speakers from Reno, Nevada, compared to Memphians.

Loud speech over noise: Some spectral attributes, with gender differences

In seeking an acoustic description of overloaded voice, simulated environmental noise was used to elicit loud speech. A total of 23 adults, 12 females and 11 males, read six passages of 90 s duration, over realistic noise presented over loudspeakers. The noise was canceled out, exposing the speech signal to analysis. Spectrum balance (SB) was defined as the level of the 2-6 kHz band relative to the 0.1-1 kHz band. SB averaged across many similar vowel segments became less negative with increasing sound pressure level (SPL), as described in the literature, but only at moderate SPL. At high SPL, SB exhibited a personal "saturation" point, above which the high-band level no longer increased faster than the overall SPL, or even stopped increasing altogether, on average at 90.3 dB (@30 cm) for females and 95.5 dB for males. Saturation occurred 6-8 dB below the personal maximum SPL, regardless of gender. The loudest productions were often characterized by a relative increase in low-frequency energy, apparently in a sharpened first formant. This suggests a change of vocal strategy when the high spectrum can rise no further. The progression of SB with SPL was characteristically different for individual subjects.

Thresholds for vowel formant discrimination using a sentence classification task

Accurate classification of vowels in sentences is challenging because American English has many acoustically similar vowels. Using a 2AFC paradigm, our previous research estimated thresholds for vowel formant discrimination in sentences that were two times smaller than the measured formant distance between close vowels. A new paradigm has been developed to estimate the ability to detect formant differences in a sentence classification task. A seven-token continuum of changes in either F1 or F2 was generated from natural productions of ‘‘bid’’ and ‘‘bed’’ using the synthesizer STRAIGHT. These tokens were spliced into a nine-word sentence at different positions that also contained two other test words, one each from pairs ‘‘cot/cut’’ and ‘‘hack/hawk.’’ Listeners were asked to identify the three words they heard in the sentence. Listeners also identified whether ‘‘bid’’ or ‘‘bed’’ was heard when only the isolated tokens were presented. Thresholds to detect a change from ‘‘bid’’ were obtained from psychometric functions fit to the data. Thresholds were similar for the sentence and word-only tasks. Overall, thresholds in both classification tasks were worse than those from the 2AFC tasks. Results will be discussed in terms of the relation between these discrimination thresholds, vowel identification, and vowel spaces. [Work supported by NIHDCD-02229.]

정보통신 단말기를 위한 한글 모음 입력 시스템

There is a limitation of using a small number of input buttons for writing Hangul words on hand-held devices. As a quick and convenient way of implementing Hangul vowels by small number of buttons, we propose a vowel input system in which vowels are fabricated from eight vowels. Our input system supports a fast input speed by making all the diphthong from one or two strokes. It also adopts a multiple input method for diphthong that users can make a diphthong in a user-friendly way of vowel writing formation or pronunciation similarity. Furthermore, we added an error correction functionality for the similar vowels that are caused by vowel harmony rules. When the proposed method is compared to the previous ones, our method outperformed in the input speed and error correction.

Neurophysiological Indexes of Speech Processing Deficits in Children with Specific Language Impairment

We used neurophysiological and behavioral measures to examine whether children with specific language impairment (SLI) have deficits in automatic processing of brief, phonetically similar vowels, and whether attention plays a role in such deficits. The neurophysiological measure mismatch negativity (MMN) was used as an index of discrimination in two tasks; one in which children ignored the auditory stimuli and watched a silent video and a second in which they attended to the auditory modality. Children with SLI showed good behavioral discrimination, but significantly poorer behavioral identification of the brief vowels than the children with typical language development (TLD). For the TLD children, two neurophysiological measures (MMN and a later negativity, LN) indexed discrimination of the vowels in both tasks. In contrast, only the LN was elicited in either task for the SLI group. We did not see a direct correspondence between the absence of MMN and poor behavioral performance in the children with SLI. This pattern of findings indicates that children with SLI have speech perception deficiencies, although the underlying cause may vary.

Discrimination and identification of long vowels in children with typical language development and specific language impairment

Researchers have claimed that children with specific language impairment (SLI) have particular difficulties in discriminating and identifying phonetically similar and brief speech sounds (Stark and Heinz, 1966; Studdert-Kennedy and Bradley, 1997; Sussman, 1993). In a recent study (Shafer et al., 2004), children with SLI were reported to have difficulty in processing brief (50 ms), phonetically similar vowels (/I-E/). The current study investigated perception of long (250 ms), phonetically similar vowels (/I-E/) in 8- to 10-year-old children with SLI and typical language development (TLD). The purpose was to examine whether phonetic similarity in vowels leads to poorer speech-perception in the SLI group. Behavioral and electrophysiological methods were employed to examine discrimination and identification of a nine-step vowel continuum from /I/ to /E/. Similar performances in discrimination were found for both groups, indicating that lengthening vowel duration indeed improves discrimination of phonetically similar vowels. However, these children with SLI showed poor behavioral identification, demonstrating that phonetic similarity of speech sounds, irrespective of their duration, contribute to the speech perception difficulty observed in SLI population. These findings suggest that the deficit in these children with SLI is at the level of working memory or long term memory representation of speech.

Cortical representation of vowels reflects acoustic dissimilarity determined by formant frequencies

We studied neuromagnetic correlates of the processing of German vowels [a], [e] and [i]. The aim was (i) to show an influence of acoustic/phonetic features on timing and mapping of the N100m component and (ii) to demonstrate the retest reliability of these parameters. To assess the spatial configuration of the N100m generators, Euclidean distances between vowel sources were computed. Latency, amplitude, and source locations of the N100m component differed between vowels. The acoustically most dissimilar vowels [a] and [i] showed more distant source locations than the more similar vowels [e] and [i]. This pattern of results was reliably found in a second experimental session after at least 5 days. The results suggest the preservation of spectral dissimilarities as mapped in a F 1– F 2 vowel space in a cortical representation.

Syllable Omission by Two-Year-Old Children

An eye-gaze and production probe examined the ability of 2-year-old children to perceive and produce minimal pairs of novel trisyllabic words with primary stress on the first or second syllables. The syllables contained dissimilar or similar vowel contrasts to determine if segments affected omission. The results supported a prosodic model, with the children's bias toward a strong-weak-weak prosodic structure and omission more frequent for the first syllable of the weak-strong-weak word pairs. Syllable omission was less frequent for strong-weak-weak word pairs that contained dissimilar vowels, suggesting that segments play a role in syllable omission.

Speech sounds learned by sleeping newborns.

It is not yet clear whether humans are able to learn while they are sleeping. Here we show that full-term human newborns can be taught to discriminate between similar vowel sounds when they are fast asleep. It is possible that such sleep training soon after birth could find application in clinical or educational situations.

Constraints of vowels and consonants on lexical selection: Cross-linguistic comparisons

Languages differ in the constitution of their phonemic repertoire and in the relative distinctiveness of phonemes within the repertoire. In the present study, we asked whether such differences constrain spoken-word recognition, via two word reconstruction experiments, in which listeners turned non-words into real words by changing single sounds. The experiments were carried out in Dutch (which has a relatively balanced vowel-consonant ratio and many similar vowels) and in Spanish (which has many more consonants than vowels and high distinctiveness among the vowels). Both Dutch and Spanish listeners responded significantly faster and more accurately when required to change vowels as opposed to consonants; when allowed to change any phoneme, they more often altered vowels than consonants. Vowel information thus appears to constrain lexical selection less tightly (allow more potential candidates) than does consonant information, independent of language-specific phoneme repertoire and of relative distinctiveness of vowels.

Pre-attentive detection of vowel contrasts utilizes both phonetic and auditory memory representations

Event-related brain potentials (ERP) were recorded to infrequent changes of a synthesized vowel (standard) to another vowel (deviant) in speakers of Hungarian and Finnish language, which are remotely related to each other with rather similar vowel systems. Both language groups were presented with identical stimuli. One standard-deviant pair represented an across-vowel category contrast in Hungarian, but a within-category contrast in Finnish, with the other pair having the reversed role in the two languages. Both within- and across-category contrasts elicited the mismatch negativity (MMN) ERP component in the native speakers of either language. The MMN amplitude was larger in across- than within-category contrasts in both language groups. These results suggest that the pre-attentive change-detection process generating the MMN utilized both auditory (sensory) and phonetic (categorical) representations of the test vowels.

Positional faithfulness, positional neutralisation and Shona vowel harmony

The distribution of the feature [high] in Shona verbs is a prototypical example of positional neutralisation accompanied by vowel harmony. In languages which exhibit positional neutralisation of vowel contrasts, one or more vowels (generally, the most marked members of the vowel inventory) may occur distinctively in only a small subset of the structural positions available in the language. Outside of these positions, the marked vowels may surface only if they harmonise with a similar vowel in the privileged position. For example, the mid vowels e and o in Shona verbs are contrastive only in root-initial syllables. These vowels may appear in subsequent syllables only when preceded by a mid vowel in root-initial position. A string of height-harmonic Shona vowels is therefore firmly anchored in the root-initial syllable, as shown in (1):