Vowel Space Research Articles

마비말장애 화자의 포먼트 전이와 모음공간에 대한 음향학적 특성

목적 : 마비말장애의 구어 명료도 분석을 위한 객관적인 음향학적 평가로 모음공간(vowel space) 분석이 유용하게 사용되고 있다. 하지만 모음공간 분석의 경우 사용자가 특정화된 수식을 이용하여 계산해야하는 수작업 측정 방식(hand-corrected measurement)으로 임상적 유용성이 낮은 단점이 있다. 이에 본 연구의 목적은 자동화 측정방식인 포먼트 전이 분석을 통해 경직형 마비말장애의 구어 명료도 측정에 대한 유용성과 모음공간 분석과의 상관관계를 알아보는 것이다.<BR>방법 : 본 연구에서는 경직형 마비말장애(n=15)와 정상 집단(n=15)을 대상으로 모음공간과 포먼트 전이 변수를 측정하였다. 모음공간은 /a/, /i/, /u/를 녹음하여 VAI, FCR, F2<SUB>/iu/</SUB> ratio를 분석하였다. 포먼트 전이는 /iu/를 빠르게 반복 시켜 F2magn, F2rate, F2reg를 분석하였다.<BR>결과 : 모음공간 분석에서 마비말장애가 정상군에 비해 VAI는 낮았고, FCR은 높았으며 F2/iu/ ratio는 통계적으로 유의하게 낮았다(p<.05). 포먼트 전이 분석에서는 마비말장애에가 정상군에 비해 F2magn, F2rate, F2reg 모두 통계적으로 유의하게 낮게 나타났다(p<.05). 모음공간과 포먼트 전이 간의 상관관계 분석을 실시한 결과 두 파라미터들 간의 높은 상관관계(r>.8)가 있었다. 이는 포먼트 전이특성인 F2magn, F2rate, F2reg는 VAI와 FCR과 높은 상관관계가 있으며 모음의 중앙화 현상으로 혀의 움직임이 저하되어 구어 명료도가 떨어짐을 알 수 있었다.<BR>결론 : 모음공간 관련 파라미터와 포먼트 전이 관련 파라미터의 객관적인 데이터를 사용하여 경직형 마비말장애의 구어 명료도 객관적으로 측정할 수 있었다. 또한 자동화된 프로그램을 사용한 포먼트 전이 분석의 경우 모음공간 분석과 높은 상관관계가 있음이 확인되어 임상에서 유용하게 사용할 수 있을 것으로 기대한다.

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).

Franklin D. Roosevelt and the American Theatre Standard: The low vowels

Franklin D. Roosevelt’s accent is often used as an illustration of the elite pronunciation heard among the north-eastern US upper classes until roughly the mid 20th century. Known under several names and often considered a mixture of British and American features, this variety is frequently identified with the American Theatre Standard, a norm popularized by acting schools in the early 20th century. Working on the assumption that Roosevelt is an exemplar of the north-eastern standard, the aim of the current study is a preliminary acoustic exploration of his accent with the aim of assessing the plausibility of such comparisons, focusing on the dress, trap, bath, start and lot vowels. Density plots created based on F1 and F2 measured in eight radio speeches were used to examine the relative position of these vowels in the vowel space. Linear mixed-effects regression was then used to model F1 and F2 in selected pairs of vowels to determine whether the differences along the two formant dimensions are significant. The results confirm a conclusion reached in an earlier auditory study (Brandenburg, Braden 1952) according to which Roosevelt’s bath was variable between [æ] and a lower and retracted [a], a vowel quality found in Eastern New England and in American Theatre Pronunciation. At the same time, a merged start/lot vowel in Roosevelt’s speech makes it unjustified to fully identify his accent with the latter variety.

A study on the change of Korean vowel sound qualities within the entire vowel space in accordance with the difference of the levels of Chinese and Vietnamese Korean learners.

A study on the change of Korean vowel sound qualities within the entire vowel space in accordance with the difference of the levels of Chinese and Vietnamese Korean learners.

Language learning as a function of infant directed speech (IDS) in Spanish: Testing neural commitment using the positive-MMR

Spanish-English bilingual families (N = 17) were recruited to assess the association between infant directed speech (IDS) in Spanish and their degree of neural commitment to the Spanish language. IDS was assessed by extracting the caregivers’ Vowel Space Area (VSA) from recordings of a storybook reading task done at home. Infants’ neural commitment was assessed by extracting the positive mismatch brain response (positive-MMR), an Event-Related Potential (ERP) thought to be indicative of higher attentional processes and early language commitment. A linear mixed model analysis demonstrated that caregivers’ VSA predicted the amplitude of the positive-MMR in response to a native speech contrast (Spanish), but not to a non-native speech contrast (Chinese), even after holding other predictors constant (i.e., socioeconomic status, infants’ age, and fundamental frequency). Our findings provide support to the view that quality of language exposure fosters language learning, and that this beneficial relationship expands to the bilingual population.

A Comprehensive Analysis of Age and Gender Effects in European Portuguese Oral Vowels

The knowledge about the age effects in speech acoustics is still disperse and incomplete. This study extends the analyses of the effects of age and gender on acoustics of European Portuguese (EP) oral vowels, in order to complement initial studies with limited sets of acoustic parameters, and to further investigate unclear or inconsistent results. A database of EP vowels produced by a group of 113 adults, aged between 35 and 97, was used. Duration, fundamental frequency (f0), formant frequencies (F1 to F3), and a selection of vowel space metrics (F1 and F2 range ratios, vowel articulation index [VAI] and formant centralization ratio [FCR]) were analyzed. To avoid the arguable division into age groups, the analyses considered age as a continuous variable. The most relevant age-related results included: vowel duration increase in both genders; a general tendency to formant frequencies decrease for females; changes that were consistent with vowel centralization for males, confirmed by the vowel space acoustic indexes; and no evidence of F3 decrease with age, in both genders. This study has contributed to knowledge on aging speech, providing new information for an additional language. The results corroborated that acoustic characteristics of speech change with age and present different patterns between genders.

Acoustic differences between voiced and whispered speech in gender diverse speakers.

Whispered speech is a naturally produced mode of communication that lacks a fundamental frequency. Several other acoustic differences exist between whispered and voiced speech, such as speaking rate (measured as segment duration) and formant frequencies. Previous research has shown that listeners are less accurate at identifying linguistic information (e.g., identifying a speech sound) and speaker information (e.g., reporting speaker gender) from whispered speech. To further explore differences between voiced and whispered speech, acoustic differences were examined across three datasets (hVd, sVd, and ʃVd) and three speaker groups (ciswomen, transwomen, cismen). Consistent with previous studies, vowel duration was generally longer in whispered speech and formant frequencies were shifted higher, although the magnitude of these differences depended on vowel and gender. Despite the increase in duration, the acoustic vowel space area (measured either with a vowel quadrilateral or with a convex hull) was smaller in the whispered speech, suggesting that larger vowel space areas are not an automatic consequence of a lengthened articulation. Overall, these findings are consistent with previous literature showing acoustic differences between voiced and whispered speech beyond the articulatory change of eliminating fundamental frequency.

(Divergent) Participation in the California Vowel Shift by Korean Americans in Southern California

This study investigates the participation in the California Vowel Shift by Korean Americans in Los Angeles. Five groups of subjects participated in a picture narrative task: first-, 1.5-, and second-generation Korean Americans, Anglo-Californians, and (non-immigrant) Korean late learners of English. Results showed a clear distinction between early vs. late bilinguals; while the first-generation Korean Americans and the late learners showed apparent signs of Korean influence, the 1.5- and the second-generation Korean Americans participated in most patterns of the California Vowel Shift. However, divergence from the Anglo-Californians was observed in early bilinguals’ speech. Similar to the late bilinguals, the 1.5-generation speakers did not systematically distinguish prenasal and non-prenasal /æ/. The second-generation speakers demonstrated a split-/æ/ system, but it was less pronounced than for the Anglo-Californians. These findings suggest that age of arrival has a strong effect on immigrant minority speakers’ participation in local sound change. In the case of the second-generation Korean Americans, certain patterns of the California Vowel Shift were even more pronounced than for the Anglo-Californians (i.e., /ɪ/-lowering, /ɑ/-/ɔ/ merger, /ʊ/- and /ʌ/-fronting). Moreover, the entire vowel space of the second-generation Korean Americans, especially female speakers, was more fronted than that of the Anglo-Californians. These findings suggest that second-generation Korean Americans may be in a more advanced stage of the California Vowel Shift than Anglo-Californians or the California Vowel Shift is on a different trajectory for these speakers. Possible explanations in relation to second-generation Korean Americans’ intersecting gender, ethnic, and racial identities, and suggestions for future research are discussed.

Diva Diction

As a follow-up to the author’s 2018 analysis of New York City English in film, this article turns its attention to the whole country over the same 80-year period of 1930–2010, using acoustic phonetic, quantitative, and statistical analysis to identify the most important changes in the pronunciation of North American English by 40 European American leading actresses in their best-known films. Focusing mostly on vowel production, the analysis reveals a gradual shift from East Coast patterns rooted in the speech of New York City to West Coast patterns rooted in the speech of Los Angeles. Changes include a decline in /r/ vocalization, which is restricted almost entirely to the period before the mid-1960s; a decline in the low back distinction between /o/ and /oh/ (lot and thought); a new distinction between /æ/ (trap) and its allophone before nasal consonants (e.g., ham or hand); shifts of /æ/ and /oh/ to a lower, more central position in the vowel space; and fronting of the back upgliding vowel /uw/ (goose). These and other patterns correspond closely to those identified in the speech of ordinary people, revealing an intriguing parallel between public speech in the mass media and private speech in local communities.

Vowels in Turkish within the framework of theories related to vowels

Dilbilim; ses, yapı ve anlam üçgeni içerisinde dilin bütünlerini ayrıntılarıyla incelemektedir. İncelemenin hareket noktası ses bölümüdür. Seslerin oluşturduğu anlamlı bütünler, yapısal birliktelikler oluşturarak üst yapıları ve anlamları meydana getirmektedir. Ses yapıları, belli kuram ve çalışmalar ışığında dilin hareketli yapısına eş bir biçimde bazen art zamanlı bazen de eş zamanlı olarak ele alınır. Çalışma, ses yapılarını ünlü uzamları esasında geçmişten günümüze belirlenen kuramlar çerçevesinde inceleme amacı taşımaktadır. Geçmişten günümüze bu konudaki ilk bu konudaki tespitlerin işitsel sesbilgisi ışığında gerçekleştiği bilinirken günümüzde ise teknoloji araçlarının sunduğu olanaklarla eklemleme ve işitsel çözümleme ile daha ayrıntılı sonuçlar elde edilebilmektedir. Ses ve ses yapıları dilbilim içerisinde “sesbilgisi (fonetik) ve sesbilim (fonoloji)” başlıkları altında incelenir. Dünya üzerinde konuşulan tüm dillerde ünlüler bulunur. Her bir dil sistemi içerisinde bu ünlülerin dağılımı ve sayısı birbirinden farklılıklar gösterir. Bunun ayırdına varan dilbilimciler özelde de sesbilimciler dünya dillerindeki ünlü yapılarını hem o dil içinde hem de diğer dillerle ilişkileri bakımından ele almıştır. Çıkan sonuçlar, bazı kuramların gelişmesini sağlayarak dil bütününün en temel verisi olan ünlüler için bizlere fikirler vermiştir. Çalışmada bu kuramların en bilinenleri ele alınarak Türkçedeki ünlü uzamı bu kuramlar çerçevesinde değerlendirilmiştir. Bununla birlikte Türkçenin ünlü nitelikleri kadın ve erkek konuşurlara ait ses laboratuvarında alınan kayıtlar ışığında formant değerleriyle çalışmada sunulmuştur. Böylelikle Türkçenin ünlü uzamının nitelik bakımından görünümü çalışmayla vurgulanmaktadır.

The Dynamics of Vowel Hypo- and Hyperarticulation in Swedish Infant-Directed Speech to 12-Month-Olds

Vowel hypo- and hyperarticulation (VHH) was investigated in Swedish infant-directed speech (IDS) to Swedish 12-month-olds using a measure that normalizes across speakers and vowels: the vhh-index. The vhh-index gives the degree of VHH for each individual vowel token, which allows for analysis of the dynamics of VHH within a conversation. Using both the vhh-index and traditional measures of VHH, the degree of VHH was compared between Swedish IDS and ADS. The vowel space area was larger in IDS than in ADS, and the average vhh-index as well as the modal value was higher in IDS than in ADS. Further, the proportion of vowel tokens that were highly hyperarticulated (vhh-index > 75th percentile) were fewer in ADS than in IDS. Vowels in Swedish IDS to 12-month-olds are thus concluded to be hyperarticulated compared to vowels in Swedish ADS, both in terms of degree and frequency. Findings are in line with previous reports on VHH in Swedish IDS as well as on VHH in IDS to infants around 12 months in other languages. The study considers the importance of robust formant estimation, highlights the need for replication of studies on VHH in IDS on previously studied languages and ages, and discusses the benefits of the vhh-index. Those benefits include that it normalizes across speakers and vowels, can be used for dynamic measures within speech samples, and permits analyses on token-level.

Examining phonetic trends in the speech of transgender YouTubers

In LGBTQ research, transgender speech is understudied compared to gay and lesbian speech. Fundamental frequency (F0) and vowel space have both shown as useful in determining speaker gender (Simpson, 2009; Gaudio, 1994). Sibilants are also used to determine gender and sexuality (Zimman, 2017; Munson, 2007). This analysis of transgender YouTubers examines F0, vowel space, and sibilants and how their speech aligns with cisgender speakers. Videos of seven transgender YouTubers were transcribed and force-aligned. Measurements were taken comparing speakers within each group and with cisgender speakers. All transgender speakers produced F0 commensurate with cisgender speakers. Vowel space was plotted, showing transwomen produce an expanded vowel space when compared to transmen. Direct comparison to cisgender speakers was not undertaken. Differences in sibilant duration between groups was not borne out in the data. Spectral mean was also measured and transwomen showed ranges similar to ciswomen. Transmen were more variable in their production of sibilants. This study shows that transgender speakers produce speech that mostly aligns with cisgender speakers for the sounds measured. Though not representative of all trans people, these speakers show gendered speech is not simply physiological and further study may aid in efforts toward greater transgender equality.

Acoustic correlates of English clear speech produced by native Korean speakers

Monolingual clear speech is acoustically distinct from casual speech and can aid speech perception (Picheny et al., 1986; Smiljanic and Bradlow, 2005) but little is known about the acoustic properties of clear speech produced by non-native speakers. The current study aims to determine whether Korean non-native speakers of English modify their speech the same way native English speakers do when asked to speak clearly. A group of native Korean speakers residing in the United States (14 recorded to date) and a group of native speakers of English (9 recorded to date) were instructed to read a list of common English words (e.g., beat) using casual and clear speaking styles. The findings showed the effect of speaking style for every acoustic measurement: vowel duration, mean pitch, pitch range, positive and negative VOT, onset f0, and vowel space. With the exception of mean pitch, there were no significant interactions between group and speaking mode, indicating that Korean speakers implemented these acoustic parameters in L2 clear speech similarly to native speakers. Only for mean pitch, the degree of change associated with clear speech was greater for English speakers than for Korean speakers.

A three-dimensional ultrasound investigation of the uvular rhotic: German /R/ in six prevocalic contexts

The German rhotic is classified in the standard variety as a uvular trill, but has been shown to demonstrate considerable variation as a result of phonetic reduction processes. Most previous work on the German /R/ has focused on its acoustic properties or its phonological patterning in particular prosodic positions, rarely addressing vowel quality and having limited means to investigate articulation. This paper presents data from a 3D/4D ultrasound investigation of word-initial /R/ in six prevocalic contexts from two L1 German speakers. The vowel space was found to be consistent with Pouplier et al. (2004). The findings indicate that a groove along the posterior dorsum accompanies trilling and that /R/ may be susceptible to coarticulatory effects from a following vowel. Specifically, tongue body height during /R/ articulation can pattern with a following vowel, inhibiting trilling and thus resulting in a fricative, complementing the EMA findings of Schiller &amp; Mooshammer (1995). This study therefore contributes to the body of literature concerning the diversity of rhotics as a class and the variability of the German rhotic in particular.

A large-scale comparison of the intelligibility of unit-selection and deep-neural-network parametric synthetic voices generated from dysarthric speech

We report on a large-scale intelligibility study of 773 listeners who transcribed semantically unpredictable stimuli generated from 4 synthetic voices: two synthetic voices, a unit-selection voice and one based on deep neural network (DNN) parametric synthesis, from the recordings of each of two dysarthric speakers. Intelligibity was calculated as the normalized phonemic edit distance (NPED) between perceived and actual transcriptions. The DNN-based voices (NPED 0.286) were significantly more intelligible (p &amp;lt; .001) than the unit-selection voices (NPED 0.316), although there was a significant interaction (p = .034) with the structure of the synthesized sentence. Counterintuitively, the voices generated from the more severely dysarthric speaker (NPED 0.272) were signficantly more intelligible (p &amp;lt; 0.001) than the other speaker's voices (NPED 0.329). Post-hoc analysis demonstrated that while the more dysarthric speaker's speech had poorer vocal quality, was measurably slower and more variable in duration, and had a smaller vowel space, this speaker also had significantly (p &amp;gt; .001) higher formant amplitudes, narrower bandwidths, and less frequency variance; she also had more careful articulation during recording. We conclude that individual differences (some within the speaker's control) can override gross measures of dysarthria in determining synthetic voice intelligibility.

Suitability of self-recordings and video calls: Vowel formants and nasal spectra

When the COVID-19 pandemic halted in-person data collection, many linguists adopted new online technologies to replace traditional methods. These included VoIP video conferencing apps like Zoom, which allow live interaction with participants, as well as user-led options in which participants record themselves using personal computers or smartphones and then email or otherwise transfer the sound files to researchers online. This study evaluated the suitability of such recordings for phonetic analysis of vowel space configurations, mergers, and nasalization by comparing simultaneous recordings from several popular personal devices (Macbook, PC laptop, iPad, iPhone, Android phone) and video call apps (Zoom, Skype, Teams) to those taken from professional equipment (H4n field recorder, Focusrite with table-top microphone). All personal devices and apps conveyed vowel arrangements and nasalization patterns relatively faithfully (especially laptops), but absolute measurements varied, particularly for the female speaker and in the 750-1500 Hz range, which affected the locations (F1xF2) of low and back vowels and reduced nasalization measurements (A1-P0) for the female's pre-nasal vowels. Based on these results we assess the validity of remote recording using these devices and offer recommendations for best practices for collecting high fidelity acoustic phonetic data from a distance.

Quantitative acoustic metrics of vowel production in mandarin-speakers with post-stroke spastic dysarthria

ABSTRACT Impairment of vowel production in dysarthria has been highly valued. This study aimed to explore the vowel production of Mandarin-speakers with post-stroke spastic dysarthria in connected speech and to explore the influence of gender and tone on the vowel production. Multiple vowel acoustic metrics, including F1 range, F2 range, vowel space area (VSA), vowel articulation index (VAI) and formant centralization ratio (FCR), were analyzed from vowel tokens embedded in connected speech produced. The participants included 25 clients with spastic dysarthria secondary to stroke (15 males, 10 females) and 25 speakers with no history of neurological disease (15 males, 10 females). Variance analyses were conducted and the results showed that the main effects of population, gender, and tone on F2 range, VSA, VAI, and FCR were all significant. Vowel production became centralized in the clients with post-stroke spastic dysarthria. Vowel production was found to be more centralized in males compared to females. Vowels in neutral tone (T0) were the most centralized among the other tones. The quantitative acoustic metrics of F2 range, VSA, VAI, and FCR were effective in predicting vowel production in Mandarin-speaking clients with post-stroke spastic dysarthria, and hence may be used as powerful tools to assess the speech performance for this population.

Association between phonation and the vowel quadrilateral in patients with stroke: A retrospective observational study.

Articulation disorder is associated with impaired control of respiration and speech organ movement. There are many cases of dysarthria and dysphonia in stroke patients. Dysphonia adversely affects communication and social activities, and it can interfere with everyday life. The purpose of this study is to assess the association between phonation abilities and the vowel quadrilateral in stroke patients.The subjects were stroke patients with pronunciation and phonation disorders. The resonance frequency was measured for the 4 corner vowels to measure the vowel space area (VSA) and formant centralization ratio (FCR). Phonation ability was evaluated by the Dysphonia Severity Index (DSI) and maximal phonation time (MPT) through acoustic evaluation for each vowel. Pearsons correlation analysis was performed to confirm the association, and multiple linear regression analysis was performed between variables.The correlation coefficients of VSA and MPT/u/ were 0.420, VSA and MPT/i/ were 0.536, VSA and DSI/u/ were 0.392, VSA and DSI /i/ were 0.364, and FCR and DSI /i/ were −0.448. Multiple linear regression analysis showed that VSA was a factor significantly influencing MPT/u/ (β = 0.420, P = .021, R2 = 0.147), MPT/i/ (β = 0.536, P = .002, R2 = 0.262), DSI/u/ (β = 0.564, P = .045, R2 = 0.256), and DSI/i/ (β = 0.600, P = .03, R2 = 0.302).The vowel quadrilateral can be a useful tool for evaluating the phonation function of stroke patients.

Stability, reliability, and sensitivity of acoustic measures of vowel space: A comparison of vowel space area, formant centralization ratio, and vowel articulation index.

Vowel space (VS) measurements can provide objective information on formant distribution and act as a proxy for vowel production. There are a number of proposed ways to quantify vowel production clinically, including vowel space area, formant centralization ratio, and vowel articulation index (VAI). The stability, reliability, and sensitivity of three VS measurements were investigated in two experiments. Stability was explored across three inter-recording intervals and challenged in two sensitivity conditions. Data suggest that VAI is the most stable measure across 30 s, 2 h, and 4 h inter-recording intervals. VAI appears the most sensitive metric of the three measures in conditions of fatigue and noise. These analyses highlight the need for stability and sensitivity analysis when developing and validating acoustic metrics, and underscore the potential of the VAI for vowel analysis.

서울말 /ᅩ/와 /ᅮ/의 지각특성

Previous studies have confirmed that /o/ and /u/ in Seoul Korean are undergoing a merger in the F1/F2 space, especially for female speakers. As a substitute parameter for formants, it is reported that female speakers use phonation (H1-H2) differences to distinguish /o/ from /u/. This study aimed to explore whether H1-H2 values are being used as perceptual cues for /o/-/u/. A perception test was conducted with 35 college students using /o/ and /u/ spoken by 41 females, which overlap considerably in the vowel space. An acoustic analysis of 182 stimuli was also conducted to see if there is any correspondence between production and perception. The identification rate was 89% on average, 86% for /o/, and 91% for /u/. The results confirmed that when /o/ and /u/ cannot be distinguished in the F1/F2 space because they are too close, H1-H2 differences contribute significantly to the separation of the two vowels. However, in perception, this was not the case. H1-H2 values were not significantly involved in the identification process, and the formants (especially F2) were still dominant cues. The study also showed that even though H1-H2 differences are apparent in females