Consonantal Context Research Articles

Formant‐frequency trajectories as acoustic correlates to speech perception.

Formant trajectories are excellent vowel discriminants; within vowel, they are nearly constant across speaker size, age, and sex, and across consonantal contexts. However, this model assumes that formant peaks are perceptually important and that human listeners track formant‐frequency changes across time. Speech‐recognition applications have avoided formant frequencies due to the difficulty of reliable formant tracking. In addition, it is not actually known whether human listeners do indeed follow formants perceptually across time. This paper presents results from several studies that examine the relationship between changing formant frequencies and perception. Alternative perceptual representations of vowels, such as global spectral shape, are precluded by evidence that individual formant amplitudes are largely ignored in vowel perception. In addition, where other spectral properties appear to have a perceptual effect, it is because stimuli have used formants that do not change. When formants are changing, perceptual effects of spectral shape properties disappear. In terms of human formant tracking, perceptual extrapolation of a formant sweep is mostly dependent on peak frequency and not other properties related to spectral shape. This demonstrates that listeners do indeed follow formant‐frequency changes as auditory objects. Further research on formant frequency perception will be described.

Perception‐production relationship in French vowel learning in adulthood.

The present study examined the extent to which perceptual performance by American English (AE) individuals predicted their accuracy in producing second‐language (L2) Parisian French (PF) vowels. Three groups of AE participants (no, moderate, and extensive French‐language experience) participated in two perceptual tasks (categorial discrimination and perceptual assimilation), and a production (repetition) task involving PF /y‐œ‐i‐a‐u/ in bilabial /rabVpa/ and alveolar /radVta/ contexts within a phrase. Results from perception tasks correctly predicted overall production difficulties and effects of language experience and consonantal context in L2 production. Paralleling their perceptual patterns, front rounded vowel productions by AE participants were mislabeled more often as back rounded vowels than as front vowels by native‐French speakers. PF /œ/ was produced more accurately with greater L2 experience. Production accuracy of /y/ was also greater with extensive experience, a finding not expected based on...

Effects of first‐language (L1) voicing assimilation rules on the second‐language (L2) perception of English word‐final consonants by Polish and Hungarian listeners.

This study explores whether knowledge of a L1 regressive voicing assimilation rule interferes with perception of word‐final voicing contrasts in an L2. Due to a regressive voicing assimilation rule, voicing in word‐final consonants is neutralized in both Hungarian and Polish. American English (AE) maintains a voicing contrast in this same context. In a perception task, 11 native Hungarian and 12 native Polish listeners identified AE word‐final consonants (e.g., [s]‐[z]) followed by voiced versus voiceless consonants. In a separate task, they identified the same consonants with the following consonantal context removed. The Hungarian and Polish listeners performed less accurately when the following context was present than when it was removed, suggesting that word‐final perception was influenced by the interference of the L1 rule.

Spectral change in the front vowels of North American English.

Assmann and Nearey [J. Acoust. Soc. Am. 80, 1297–1308 (1986)] coined the term “vowel‐inherent spectral change” (VISC) to refer to change in spectral properties inherent to the phonetic specification of vowels. Although VISC includes the relatively large formant movements associated with acknowledged diphthongs, it was explicitly extended to include reliable (but possibly more subtle) spectral change associated with vowel categories of North American English typically designated as monphthongs. This paper reviews statistical evidence of VISC in the formant patterns in front vowels of /hVd/ syllables in three regional dialects of English: Dallas, TX [Assmann and Katz, J. Acoust. Soc. Am. 108, 1856–1866 (2000)], Western, MI [Hillenbrand et al., J. Acoust. Soc. Am. 97, 3099–3111 (1995)] and Northern, AB (new data). Results suggest that VISC patterns may be useful characteristics for assessing dialect differences. Evidence is presented for the importance of VISC in vowel perception, including recent evidence from our laboratories regarding perception by second language learners. A progress report is provided on research into how VISC is best characterized parametrically and which temporal regions of a vocoid may be most effective in summarizing VISC patterns in varying consonantal contexts.

Language experience and consonantal context effects on perceptual assimilation of French vowels by American-English learners of French

Recent research has called for an examination of perceptual assimilation patterns in second-language speech learning. This study examined the effects of language learning and consonantal context on perceptual assimilation of Parisian French (PF) front rounded vowels /y/ and /oe/ by American English (AE) learners of French. AE listeners differing in their French language experience (no experience, formal instruction, formal-plus-immersion experience) performed an assimilation task involving PF /y, oe, u, o, i, epsilon, a/ in bilabial /rabVp/ and alveolar /radVt/ contexts, presented in phrases. PF front rounded vowels were assimilated overwhelmingly to back AE vowels. For PF /oe/, assimilation patterns differed as a function of language experience and consonantal context. However, PF /y/ revealed no experience effect in alveolar context. In bilabial context, listeners with extensive experience assimilated PF /y/ to (j)u less often than listeners with no or only formal experience, a pattern predicting the poorest /u-y/ discrimination for the most experienced group. An "internal consistency" analysis indicated that responses were most consistent with extensive language experience and in bilabial context. Acoustical analysis revealed that acoustical similarities among PF vowels alone cannot explain context-specific assimilation patterns. Instead it is suggested that native-language allophonic variation influences context-specific perceptual patterns in second-language learning.

Speaker variability when producing repeated syllables and across speech tasks.

Target theory [Lindblom, J. Acoust. Soc. Am. 35, 1773–1781 (1963)] of vowel perception and production suggests that vowels have unique and characteristic articulatory and acoustic events that constitute the canonical forms (“targets”) of vowels. However, speakers do not always achieve these targets. There exists substantial variability in speech within a speaker and across speakers. Some of the challenges in understanding speech production lie in variability resulting from various factors including speaker identity, prosody, rate, and speech task. Because of the lack of data permitting good statistical estimates of target events, the following experiments were performed. First, within‐speaker variations in consonant‐vowel (CV) and vowel‐consonant transitions in three monosyllabic words (dock, knock, and shock) during a sentence reading repetition task are evaluated for two speakers. Second, acoustic vowel targets in /h/‐vowel‐/d/ or‐/t/ and selected consonant contexts (CV‐/d/ or‐/t/) are studied in the speech production of one individual across three speech tasks: citation, passage reading, and conversational speech. The hypothesis is a coarse‐grained consistency with fine‐grained variability in the acoustic signal and a continuum of changes in the acoustic vowel targets across tasks. Findings and continuing directions are discussed within the framework of target theory.

Acoustic and articulatory manifestations of vowel reduction in German

Recent phonological approaches incorporate phonetic principles in the motivation of phonological regularities, e.g. vowel reduction and neutralization in unstressed position by target undershoot. So far, evidence for this hypothesis is based on impressionistic and acoustic data but not on articulatory data. The major goal of this study is to compare formant spaces and lingual positions during the production of German vowels for combined effects of stress, accent and corrective contrast. In order to identify strategies for vowel reduction independent of speaker-specific vocal-tract anatomies and individual biomechanical properties, an approach similar to the Generalized Procrustes Analysis was applied to formant spaces and lingual vowel target positions. The data basis consists of the German stressed and unstressed full vowels /iù ɪ yù ʏ eù ɛ ɛùφùœaù a où ɔ uù ʊ/ from seven speakers recorded by means of electromagnetic midsagittal articulography (EMMA). Speaker normalized articulatory and formant spaces gave evidence for a greater degree of coarticulation with the consonant context for unstressed vowels as compared to stressed vowels. However, only for tense vowels could spatial reduction patterns be attributed to vowel shortening, whereas lax vowels were reduced without shortening. The results are discussed in the light of current theories of vowel reduction, i.e. target undershoot, Adaptive Dispersion Theory and Prominence Alignment.

Amplitude variations in coarticulated vowels

This paper seeks to characterize the nature, size, and range of acoustic amplitude variation in naturally produced coarticulated vowels in order to determine its potential contribution and relevance to vowel perception. The study is a partial replication and extension of the pioneering work by House and Fairbanks [J. Acoust. Soc. Am. 22, 105-113 (1953)], who reported large variation in vowel amplitude as a function of consonantal context. Eight American English vowels spoken by men and women were recorded in ten symmetrical CVC consonantal contexts. Acoustic amplitude measures included overall rms amplitude, amplitude of the rms peak along with its relative location in the CVC-word, and the amplitudes of individual formants F1-F4 along with their frequencies. House and Fairbanks' amplitude results were not replicated: Neither the overall rms nor the rms peak varied appreciably as a function of consonantal context. However, consonantal context was shown to affect significantly and systematically the amplitudes of individual formants at the vowel nucleus. These effects persisted in the auditory representation of the vowel signal. Auditory spectra showed that the pattern of spectral amplitude variation as a function of contextual effects may still be encoded and represented at early stages of processing by the peripheral auditory system.

Production of Parisian French front rounded vowels by second-language learners

This study examined the effects of language experience and consonantal context on the production of Parisian French (PF) vowels by American English (AE) learners of French. A repetition task was performed, involving French vowels /y-œ-i-a-u/ uttered by a native speaker of Parisian French (PF) in bilabial /bVp/ and alveolar /dVt/ contexts embedded in the nonsense word/phrase /raCVCa/. Three groups of native AE speakers participated: speakers without French experience (NoExp), speakers with formal French experience (ModExp), and speakers with formal and extensive immersion experience (HiExp). Production accuracy was assessed by native PF listeners' judgments and by acoustic analysis. Native PF listeners identified second-language (L2) learners' productions more accurately as a function of speakers' increased language experience, although /u/, /y/ and /œ/, even produced by the HiExp group, were frequently misidentified. A consonantal context effect was revealed, including /u/ being misidentified more often in alveolar context than in bilabial context by all three groups. Overall, all groups distinguished front rounded /y/ from /u/ in production, but often in a non-native manner, e.g., producing /y/ as /ju/. Individual differences, interactions between consonantal context and vowels, as well as implications for the perception-production link in L2 learning, are discussed.

Regional variation in vowels and vowel systems: normalization and optimization

We made an acoustical description of regional variation patterns in the vowel system of Dutch spoken in the Netherlands and Flanders. The speech material consisted of read monosyllabic utterances in a neutral consonantal context, representing the vowels of Dutch. A discriminant analysis applied on the raw measurements to classify each speaker into one of the eight regions involved (socio‐geographic variation) gave a correct regional classification for 72.0% of the 160 speakers. When normalization procedures were applied, the percentages of proper classification increase to 82.5%. Several questions have to be answered though. Which normalization procedure is the best one and why, and how can we be sure that specific parts of sociophonetic variation are not distorted by the normalization procedure? We will present additional materials collected on the same set of speakers in reading aloud a strictly controlled list of words containing all Dutch vowels in different consonantal contexts. Another question is how to reduce the amount of information used in a normalization procedure. The z transformation (Lobanov) performs well, but is rather expensive in the amount of information needed. Can we minimize the information needed and yet uncover the patterns of sociophonetic variation?

The temporal location of rms peak in coarticulated vowels

The acoustic energy peak of the vowel, here defined as the rms peak, occurs typically before the temporal vowel midpoint. This study examines whether and how the location of the rms peak varies as a function of consonantal context and vowel duration. Four measures of rms peak location were explored: its distance (in ms) from vowel onset and from CVC‐word onset (absolute measures) and its relative position (ranging from 0 to 100%) within the vowel and within the CVC‐word (relative measures). Analysis of eight American English vowels produced in ten consonantal contexts by twenty speakers shows that the two relative measures yielded the most comparable and consistent patterns. The relative location of rms peak varied significantly as a function of consonantal context. It occurred earlier (i.e., closer to vowel or word onset) in the context of fricatives (and not stops), voiced consonants (as opposed to voiceless), and labials (as opposed to alveolars). The temporal location of rms peak was also closer to the vowel onset (or word onset) for long vowels as opposed to short. Explanation of these patterns will be provided.

A cross‐dialect acoustic description of vowels: Peruvian versus European Spanish

This paper aims to provide acoustic description and comparison of vowel properties of two major dialects of Spanish: Peruvian and European. The method of data collection and acoustic analyses reported in Escudero et al. (submitted) for Portuguese is replicated in order to allow for future comparisons across languages. Each of the five Spanish monophthongs /i, e, a, o, u/ was produced in stressed position in the first syllable of a CVCV (consonant‐vowel‐consonant‐vowel) frame. The CVCV words were read both in isolation and in a carrier sentence. Vowel tokens were examined between voiceless stops /p, t, k/ and fricatives /f, s/, which supplied five variable consonantal contexts. We analyzed the vowels' first and second formants (F1 and F2) as well as their fundamental frequency and duration. Unsurprisingly, a main effect of vowel identity and gender on formant values was found. Most importantly, unlike Morrison and Escudero (2007), we found for Spanish an interaction between dialect and vowel identity, implying that the F1 and F2 values of some vowels are significantly different between dialects.

La distribució de les vocals mitjanes tòniques en el català central

Mid vowels in stressed position are either closed ore open in Central Catalan, depending on a number of factors. In this paper we examine and discuss several of the proposals in the literature (general tendency towards lowering, influence of the immediate right consonantal context, lowering effect by the presence of a high front vocoid in the following syllable(s), analogical effects stemming from the existence of phonologically similar elements in the lexicon, and metapaphonic effects caused by a following unstressed mid closed vowel. These independent factors cause a certain predominance of mid open vowels in cases of indeterminacy, but a general tendency toward open mid vowels based on a tendency to increase sonority is shown to be unjustified.

Perceptual similarity of American English and Japanese vowels for native speakers of American English and Japanese

Native speakers of American English and Japanese chose American English vowels that best represented Japanese vowels uttered in two different consonantal contexts. The two groups of subjects’ responses differed noticeably, showing the effects of linguistic experience. The Japanese subjects tended to match Japanese long two-mora vowels with long, tense vowels and short one-mora vowels with short lax vowels, but the English subjects were relatively unaffected by durational differences. The Japanese subjects matched /a/ and /e/ to /æ/ and /[g\/]/, respectively, but the English subjects matched /a/ with /■/ or /■/, and /e/ with /■/ as well as /[g\/]/. The Japanese subjects matched /o/ with /■/, while the American subjects matched it with /o■/. The Japanese subjects took part in two other experiments. They equated American English vowels to Japanese vowel categories and they also identified American English vowels. They equated English /■/ to Japanese /a/, and /■/ was often misidentified as /æ/. The results of the three experiments show that there are discrepancies between how an American English vowel is mapped into Japanese vowel categories and how the Japanese subjects expect it to sound, which may result in the poor identification of American English vowels.

Effects of several consonant environments on vowel formants

A large body of evidence has shown that relative change in spectral pattern possesses some invariant properties for vowels across speakers in /hVd/ environment. While formant frequency plots from vowel steady-states result in categories with large overlap, better separation of vowel categories is obtained using onset and offset formant frequencies [e.g., Hillenbrand et al., J. Acoust. Soc. Am. 97, 3099–3111 (1995)]. To a large extent, this invariance was also shown in other /CVC/ contexts using all combinations of /h,b,d,g,p,t,k/ [Hillenbrand et al., J. Acoust. Soc. Am. 109, 748–763 (2001)]. The present study explored spectral change in vowels produced in environments where larger differences might be expected across contexts. Nineteen men and 39 women were asked to produce fourteen Canadian English vowels in /hV/, /hVd/, /hVt/, /hVl/, /hVr/, /hVnd/, /hVg/, and /dVd/. Subjects read standard English orthographic representations of the target words which were embedded both in a sentence which indicated the rhyme of the target word—e.g., ‘‘Swooned rhymes with hoond’’ as well as in a sentence which indicated the pronunciation of onset and vowel—e.g., ‘‘Hood sounds like hoog.’’ Formant frequencies were measured and tracked for each token and differences across consonant contexts were analyzed. [Work supported by SSHRC.]

Dynamic spectral structure specifies vowels for children and adults.

When it comes to making decisions regarding vowel quality, adults seem to weight dynamic syllable structure more strongly than static structure, although disagreement exists over the nature of the most relevant kind of dynamic structure: spectral change intrinsic to the vowel or structure arising from movements between consonant and vowel constrictions. Results have been even less clear regarding the signal components children use in making vowel judgments. In this experiment, listeners of four different ages (adults, and 3-, 5-, and 7-year-old children) were asked to label stimuli that sounded either like steady-state vowels or like CVC syllables which sometimes had middle sections masked by coughs. Four vowel contrasts were used, crossed for type (front/back or closed/open) and consonant context (strongly or only slightly constraining of vowel tongue position). All listeners recognized vowel quality with high levels of accuracy in all conditions, but children were disproportionately hampered by strong coarticulatory effects when only steady-state formants were available. Results clarified past studies, showing that dynamic structure is critical to vowel perception for all aged listeners, but particularly for young children, and that it is the dynamic structure arising from vocal-tract movement between consonant and vowel constrictions that is most important.

Acoustic variability within and across German, French, and American English vowels: Phonetic context effects

Cross-language perception studies report influences of speech style and consonantal context on perceived similarity and discrimination of non-native vowels by inexperienced and experienced listeners. Detailed acoustic comparisons of distributions of vowels produced by native speakers of North German (NG), Parisian French (PF) and New York English (AE) in citation (di)syllables and in sentences (surrounded by labial and alveolar stops) are reported here. Results of within- and cross-language discriminant analyses reveal striking dissimilarities across languages in the spectral/temporal variation of coarticulated vowels. As expected, vocalic duration was most important in differentiating NG vowels; it did not contribute to PF vowel classification. Spectrally, NG long vowels showed little coarticulatory change, but back/low short vowels were fronted/raised in alveolar context. PF vowels showed greater coarticulatory effects overall; back and front rounded vowels were fronted, low and mid-low vowels were raised in both sentence contexts. AE mid to high back vowels were extremely fronted in alveolar contexts, with little change in mid-low and low long vowels. Cross-language discriminant analyses revealed varying patterns of spectral (dis)similarity across speech styles and consonantal contexts that could, in part, account for AE listeners' perception of German and French front rounded vowels, and "similar" mid-high to mid-low vowels.

Perception of French vowels by American English adults with and without French language experience

This study investigated the effects of language experience and consonantal context on American English (AE) listeners’ discrimination of contrasts involving Parisian French vowels /y, œ, u, i/. Vowels were produced in /rabVp/ and /rabVt/ nonsense disyllables in carrier phrases by 3 speakers and presented in a categorial AXB discrimination task. Two groups were tested: AE listeners who had studied French extensively beginning after age 13 (Exp) and non-French-speaking AE listeners (Inexp). The Exp group performed better than the Inexp group on /u-œ/, /i-y/ and /y-œ/ (mean errors: Exp=5%, Inexp=24%). However, for /u-y/, the groups did not differ (Exp=30% vs Inexp=24% errors). The Inexp group confused /i-y/ more often in bilabial context, but /u-y/ more often in alveolar context, whereas the Exp group confused /u-y/ in both contexts. Overall, the Inexp group performed better in bilabial than in alveolar context (16% vs 32% errors), whereas the Exp group revealed no context effect. Results suggest that learning a second language (L2) includes learning its coarticulatory rules. Implications for models of L2-speech perception are discussed.

The Contribution of Various Sources of Spectral Mismatch to Audible Discontinuities in a Diphone Database

One of the major problems in concatenative synthesis is the occurrence of audible discontinuities between two successive concatenative units. Several studies have attempted to discover objective distance measures that predict the audibility of these discontinuities. In this paper, we investigate mid-vowel joins for three vowels with a range of post-vocalic consonant contexts typical for diphone databases. A first perceptual experiment uses a pairwise comparison procedure to find two subsets of unit combinations: Those with versus without audible discontinuities. A second perceptual experiment uses these two subsets in a procedure where formant resynthesis is used to manipulate three sources of discontinuity separately: formant frequencies, formant bandwidths, and overall energy. Results show mismatch in formant frequencies provides the largest contribution to audible discontinuity, followed by mismatch in overall energy

An acoustic description of the vowels of northern and southern standard Dutch II: Regional varieties

An analysis is presented of regional variation patterns in the vowel system of Standard Dutch as spoken in the Netherlands (Northern Standard Dutch) and Flanders (Southern Standard Dutch). The speech material consisted of read monosyllabic utterances in a neutral consonantal context (i.e., /sVs/). The analyses were based on measurements of the duration and the frequencies of the first two formants of the vowel tokens. Recordings were made for 80 Dutch and 80 Flemish speakers, who were stratified for the social factors gender and region. These 160 speakers were distributed across four regions in the Netherlands and four regions in Flanders. Differences between regional varieties were found for duration, steady-state formant frequencies, and spectral change of formant frequencies. Variation patterns in the spectral characteristics of the long mid vowels /e o ø/ and the diphthongal vowels /ei oey bacwards c u/ were in accordance with a recent theory of pronunciation change in Standard Dutch. Finally, it was found that regional information was present in the steady-state formant frequency measurements of vowels produced by professional language users.