Interaction between duration, context, and speaking style in English stressed vowels (original) (raw)

The Acoustic Correlates of Stress and Accent in English Content and Function Words

This paper has two aims: (1) To contribute to the discussion on what the acoustic correlates of stress and accent in English are, a question on which there is currently no universal agreement; (2) To determine whether vowels in function words receive less stress than similarly unstressed vowels in content words. To this purpose, the study analyses 614 occurrences of the lax high front vowel /I/ in read speech produced by 10 male speakers of Standard Southern British English. 14 different acoustic features are investigated. Results indicate that (1) there are two acoustic correlates of accent (duration and f0 slope), four acoustic correlates of stress (spectral balance/tilt, intensity/loudness, amplitude of voicing (H1) and amplitude of the first harmonic (A1)), three acoustic correlates of prominence in general (F1, H1-A2 and H1-A3), and four acoustic features that appear to be unrelated to the expression of accent, stress or prominence (F2, HNR, glottal leakage (B1) and the open quotient (H1-H2)). Regarding question (2), there is also limited evidence that British English function words might be less prominent than unstressed syllables in content words.

STRESSED VOWEL DURATION AND PHONEMIC LENGTH CONTRAST

It has been generally accepted that greater vowel/syllable duration is a reliable correlate of stress and that absolute durational differences between vowels underlie phonemic length contrasts. In this paper we shall demonstrate that duration is not an independent stress correlate, but rather it is derivative of another stress correlate, namely pitch. Phonemic contrast, on the other hand, is qualitative rather than quantitative. These findings are based on the results of an experiment in which four speakers of SBrE read 162 mono-, di-and trisyllabic target items (made of CV sequences) both in isolation and in carrier phrases. In the stressed syllables all Southern British English vowels and diphthongs were represented and each vowel was placed in 3 consonantal contexts: (a) followed by a voiced obstruent, (b) voiceless obstruent and (c) a sonorant. Then, all vowels (both stressed and unstressed) were extracted from target items and measured with PRAAT. The results indicate that stressed vowels may be longer than unstressed ones. Their durational superiority, however, is not stress-related, but follows mainly from vowel-intrinsic durational characteristics and, to some extent, from the prosodic context (i.e. the number of following unstressed vowels) in which it is placed. In CV 1 CV 2 disyllables, when V 1 is phonemically short, the following word-final unstressed vowel is almost always longer. It is only when V 1 is a phonemically long vowel that V 2 may be shorter. As far as diphthongal V 1 is concerned, the durational V 1 ~V 2 relation is variable. Interestingly, the V 1 ~V 3 relation in trisyllables follows the same durational pattern. In both types of items the rare cases when a phonemically short V 1 is indeed longer than the word-final vowel involve a stressed vowel which is open, e.g. [], and whose minimal execution time is longer due to a more extensive jaw movement. These observations imply that both in acoustic and perceptual terms the realisation of word stress is not based on the durational superiority of stressed vowels over unstressed ones. When it is, it is only an epiphenomenon of intrinsic duration of the stressed vowel and extra shortness of non-final unstressed vowel. As far as phonemic length contrast is concerned, we observe a high degree of durational overlap between phonemically long and short vowels in monosyllabic CVC words (which is enforced by a greater pitch excursion), whereas in polysyllables the differences seem to be perceptually non-salient (>40 ms, cf. Lehiste 1970). This suggests that the differences in vowel duration are not significant enough to underlie phonological length contrasts.

Durational effects of stress, accent, and voicing on the preceding word-final syllable in English

This study examined how duration of an unstressed final syllable in English is affected by conditions in the following word: stress (trochaic/iambic), accent (accented/unaccented), and initial stop voicing (voiced/voiceless). Results showed that the unstressed final syllable was shorter before an unstressed syllable, presumably due to polysyllabic shortening-i.e., the following unstressed syllable forms a foot with the preceding syllable. This effect, however, disappeared when the following word was accented, due to foot restructuring caused by leftward spreading of accent effect-i.e., because the (following) unstressed syllable is lengthened when accented, it is no longer weak enough to be associated with the preceding foot. The lengthening of the word-final syllable before a voiced stop was also observed, but only within a foot. Most of the foot restructuring effects disappeared across an IP boundary. Interestingly, however, even across an IP boundary, the final syllable was affected by accentuation of the following word (i.e., shortened before an accented word), implying that the prominence structure may have a more global effect.

Cross-linguistic differences between accented vs unaccented vowel durations

This study compares durational measures of accented vs unaccented vowels for data of 5 languages (4 speakers per language read translations of the same passage). Results show clear cross-language differences: accented vowels in our data of British English are more marked in terms of duration than in data of Mexican Spanish and Parisian French (data of German and Italian take a somewhat intermediate position). Direct durational measures of accented vs unaccented vowels yield a useful insight in aspects of speech rhythm that are only partially addressed by global measures of V variability (as provided by various popular rhythm metrics). Such results may have implications in rhythm typology and might help understand why stresses are perceived to occur at roughly regular intervals of time in so-called stress-timed languages.

Acoustic Correlates of Word Stress as A Cue to Accent Strength

Research in Language, 2014

Due to the clear interference of their mother tongue prosody, many Czech learners produce their English with a conspicuous foreign accent. The goal of the present study is to investigate the acoustic cues that differentiate stressed and unstressed syllabic nuclei and identify individual details concerning their contribution to the specific sound of Czech English. Speech production of sixteen female non-professional Czech and British speakers was analysed with the sounds segmented on a word and phone level and with both canonical and actual stress positions manually marked. Prior to analyses the strength of the foreign accent was assessed in a perception test. Subsequently, stressed and unstressed vowels were measured with respect to their duration, amplitude, fundamental frequency and spectral slope. Our results show that, in general, Czech speakers use much less acoustic marking of stress than the British subjects. The difference is most prominent in the domains of fundamental freq...

Effect of speaking rate and contrastive stress on formant dynamics and vowel perception

The Journal of the Acoustical Society of America, 2000

Vowel formants play an important role in speech theories and applications, however, the same formant values measured for the steady-state part of a vowel can correspond to different vowel categories. Experimental evidence indicates that dynamic information can also contribute to vowel characterization. Hence dynamically modeling formant transitions may lead to quantitatively testable predictions in vowel categorization. Because the articulatory strategy used to manage different speaking rates and contrastive stress may depend on speaker and situation, the parameter values of a dynamic formant model may vary with speaking rate and stress. In most experiments speaking rate is rarely controlled, only 2 or 3 rates are tested, and most corpora contain just a few repetitions of each item. As a consequence, the dependence of dynamic models on those factors is difficult to gauge. This article presents a study of 2,300 [iai] or [iεi] stimuli produced by two speakers at 9 or 10 speaking rates in a carrier sentence for 2 contrastive stress patterns. The corpus was perceptually evaluated by naive listeners. Formant frequencies 1 were measured during the steady-state parts of the stimuli, and the formant transitions were dynamically and kinematically modeled. The results indicate that (1) the corpus was characterized by a contextual assimilation instead of a centralization effect; (2) dynamic or kinematic modeling was equivalent as far as the analysis of the model parameters was concerned; (3) the dependence of the model parameter estimates on speaking rate and stress suggests that the formant transitions were sharper for high speaking rate, but no consistent trend was found for contrastive stress; (4) the formant frequencies measured in the steady-state parts of the vowels were sufficient to explain the perceptual results while the dynamic parameters of the models were not.

Acoustic Properties of Japanese and English Vowels: Effects of Phonetic and Prosodic Context

search has documented large differences in the phonetic realization of American English (AE) vowels as a function of phonetic context and prosodic context Strange et al., submitted). Less research has been published on the allophonic and prosodic variation of Japanese (J) vowels. Thus, the goal of the present study was to compare the allophonic and prosodic variation in spectral and temporal structure of J and AE vowels, using acoustical analysis of corpora in which the phonetic and prosodic context was varied systematically. To the extent that the type and amount of phonetic variation differs across the two languages, we would expect that cross-language perceptual similarity might also vary with contextual variables.

The pronunciation of vowels with secondary stress in English

2018

There are few studies which have focused on the pronunciation of vowels with secondary stress in English. Within the framework introduced by Guierre (1979), this paper offers a large empirical study of these vowels and focuses on three key categories of words: non-derived words, constructions containing a semantically transparent prefix and suffixal derivatives. Overall, previous analyses based on rank, phonological domains and derivational isomorphism are confirmed but certain phenomena unveiled by this study require a few revisions of existing models. Several possible formalisation options are proposed to implement these revisions. The most promising seems to be the (here, exploratory) analysis using feet because it could considerably broaden the empirical coverage of the theory.