Series Of Vowels Research Articles

Formant frequencies, vowel identity, and the perceived relative tallness of synthetic speakers

Listeners can make consistent judgments regarding the tallness of speakers [Rendall etal., J. Exp. Psych: Human Percep. Perform. 33, 1208 (2007)]. These judgments are informed by the f0 and formant frequencies (FFs) of a speaker's voice. However, FFs are also cues to vowel identity, such that a small speaker producing an /u/ might have lower average FFs than a larger speaker producing an /æ/. Do listeners use absolute FFs to judge speaker tallness, or do they “correct” for phonetic identity and consider the FFs of a vowel relative to those expected for that category? To test this, a series of synthetic vowels (/i æ u/) with different FF scalings and different numbers of formants (2–5) were created. These scalings were intended to replicate speakers of different vocal tract lengths (i.e., sizes). Participants were presented with pairs of vowels and asked to indicate which vowel sounded like it had been produced by a taller speaker. Results indicate that listeners consider both absolute and phonetically “corrected” FF information and that formants higher than F3 greatly reduce listener's reliance on absolute F1 and F2 information in making speaker size judgments.

Cross-linguistic differences in the immediate serial recall of consonants versus vowels

ABSTRACTThe current study investigated native English and native Arabic speakers’ phonological short-term memory for sequences of consonants and vowels. Phonological short-term memory was assessed in immediate serial recall tasks conducted in Arabic and English for both groups. Participants (n= 39) heard series of six consonant–vowel syllables and wrote down what they recalled. Native speakers of English recalled the vowel series better than consonant series in English and in Arabic, which was not true of native Arabic speakers. An analysis of variance showed that there was an interaction between first language and phoneme type. The results are discussed in light of current research on consonant and vowel processing.

Perceptual normalization for variation in speaking style.

Ladefoged and Broadbent [J. Acoust. Soc. Am. 29, 98–104 (1957)] demonstrated that listeners will shift their categorization of a target vowel as a function of acoustic characteristics of a preceding carrier phrase. These results have been interpreted as an example of perceptual normalization for variability resulting from differences in talker anatomy. The present study examined whether listeners would normalize for acoustic variability resulting from differences in speaking style within a single talker. Two vowel series were synthesized that varied between central and peripheral vowels (the vowels in “beat”‐“bit” and “bod”‐“bud”). Each member of the series was appended to one of four carrier phrases that were spoken in either a “clear” or “reduced” speech style. Participants categorized vowels in these eight contexts. A reliable shift in categorization as a function of speaking style was obtained for three of four phrase sets. This demonstrates that phrase context effects can be obtained with a single talker. However, the directions of the obtained shifts are not reliably predicted on the basis of the speaking style of the talker. Instead, it appears that the effect is determined by an interaction of the average spectrum of the phrase with the target vowel. [Work supported by NIH‐NIDCD.]

Does training improve consistency of roughness judgments in a matching task?

Shrivastav et al. [J. Acoust. Soc. Am. 119(5), 3340 (2006)] reported a matching task in which a square wave modulated sawtooth wave (signal) was matched to a series of vowels (standards) to estimate the magnitude of roughness in voices. Results suggested that listeners found it difficult to isolate roughness from other voice quality dimensions such as breathiness. In the present experiment, a brief training session was added prior to the matching task to ensure that listeners were attending to variations in roughness alone. This training consisted of a rank‐ordering task in which listeners sorted the 34 test samples of the vowel /a/ in ascending order of roughness. For feedback, these rankings were compared to the rankings made by expert listeners. This criterion was used to ensure that all participants understood the dimension of voice quality under study. Listeners who met a specific eligibility criterion completed a modulation matching task to familiarize them with the matching task itself. Finally, thresholds for the roughness matching task were obtained and compared to thresholds obtained without training. The extent to which task‐related training can help listeners make perceptual judgments for specific dimensions of voice quality will be discussed.

Native language experience influences the perceived similarity of second language vowel categories.

Most studies of L2 speech perception seek to characterize—at least implicitly—how the similarity among L2 speech sound categories is shaped by L1 experience. Intercategory similarity for native language speakers is rarely considered, however. Here, we derive two indices of graded intercategory similarity for a front vowel series (pin, pen, and pan): One from an off‐line perceptual similarity judgment task and a second from online measures of arm‐movement trajectories in a word recognition task. Both tasks revealed graded effects of intercategory similarity, but the similarity spaces differed between language groups. Both groups perceived /ae/ and /E/ to be most similar, but the native Italian speakers perceived /I/ to be equally similar to /E/ and /ae/, whereas English speakers perceived the /I/ and /E/ to be more similar than /I/ and /ae/. The Italian speakers’ performance in both tasks suggests a similarity space that is dominated by dimensions that are contrastive for /I/ with respect to the other vowels. This may reflect a compensatory strategy for L2 vowel perception that may derive from experience with the distributional properties of their native vowels.

National and ethnic identity markers

The New Zealand (NZ) short front vowels are often considered as a defining feature of New Zealand English (NZE), yet research which has considered data from both the Pakeha (NZ European) and the NZ Maori communities has noted slightly different patterns in the realisations of the vowel in the KIT lexical set in the respective communities (Bell 1997a, b; Warren and Bauer 2004). This paper compares the short front vowel series of NZ Maori students with that of NZ Samoan, Tongan, Cook Island and Niuean students and demonstrates how the NZ short front vowel series mark both similarity and difference across NZ communities. Our findings show that NZ Maori students have a greater degree of centralisation in their KIT vowel and a greater degree of raising of their DRESS and TRAP vowels than their NZ Pasifika counterparts. However, the manner in which the vowels raise and centralise distinguishes NZ Maori and Cook Island students from their NZ Samoan, Tongan and Niuean cohorts. The latter observation highlights problems with the pan-ethnic “Pasifika” label used to distinguish NZ Maori from other NZ Polynesian communities.

Salience of dynamic virtual formants in diphthongs

Variations in the quality of a synthesized vowel can be produced by changing the frequency of its steady-state formants or the amplitude of closely-spaced static spectral components—the latter effect a function of auditory spectral integration. Recent studies have demonstrated that dynamic modification of the amplitude ratios of these spectral components can give rise to the perception of virtual diphthongs. This study examines the extent to which spectral integration of these dynamic components is uniform across a range of frequency separations of these spectral components (previous work [Fox et al., ‘‘Title,’’ J. Acoust. Soc. Am. 120, 3252 (2006)] demonstrated the integration of static components beyond 3.5 bark). Two synthetic vowel series were created by modifying the rise or fall of F2 producing [■i]-[■:]-[■u] and [i■]-[■:]-[u■] series. Next, several different series were created in which this dynamic F2 was replaced by two static spectral components whose amplitude ratios were varied over time (virtual F2 glides). The spectral separation of these two components varied from 2.0 to 4.0 bark. A single-interval 2AFC identification task was used to obtain estimates of vowel quality. Results will be discussed in terms of whether the spectral integration effects decline as the frequency separation between components increases. [Work supported by NIDCD R01DC00679-01A1.]

Die Kehlkopfkonfiguration beim Flüstern

The purpose of the study was to assess glottal and supraglottal configurations during whispering in comparison to habitual phonation and to visually estimate to what extent force is exerted onto the vocal fold epithelium during whispering. Is whispering a good compromise between preservation of epithelium and permission to communicate orally, e. g. postoperatively? We investigated on 100 patients with organic or non-organic voice disorders. Patients with vocal fold immobility and patients with organic lesions large enough to impede glottal closure were excluded. Videolaryngoscopy and stroboscopy was performed via flexible endoscopy. During fiberoscopy the patients were asked (i) to phonate and then whisper a series of vowels in various loudness levels and (ii) to count to 10 voiced and whispered in various loudness levels. The extent of vocal fold adduction and supraglottal configuration were assessed and whispering compared with voiced phonation. Interindividual differences in glottal and supraglottal configuration during whispering were seen. The two different and most frequently seen glottal area patterns were determined by the position of the vocal processes - either adducted or abducted. Adduction resulted in sole cartilaginous triangle formation or 'kissing processes', abduction in ligamentous and cartilaginous triangle. Another surprising result was that with increasing loudness either a narrowing or widening of glottal and supraglottal structures was visible. There are different patterns of glottal and supraglottal configuration during whispering. Although the vocal folds can show partial or total contact over the membranous parts, the lack of additional vibration of the vocal folds suggests that epithelial forces are smaller than with additional vibratory shear stress. Thus, whispering may be considered and allowed for postoperative communication purposes, when patients follow instructions for low-effort whispering.

Salience of virtual formants as a function of the frequency separation between spectral components

The center-of-gravity (COG) hypothesis proposed by Chistovich and others for the perception of static vowels suggests that auditory spectral integration may occur when two or more formants fall within a 3.5 bark bandwidth. While several studies have examined the bandwidth limits of such integration, this study examines the extent to which spectral integration is uniform within this putative 3.5-bark range. We examine the perceptual salience of virtual formants produced by modifying the spectral COG of two closely spaced narrow-bandwidth resonances. Three different vowel series were created: [i-■], [■-■] and [■-■]. A second set of vowels was then created in which one of the formants (F1 in [i-■], F2 in [■-■] and F3 in [■-■]) was replaced by a virtual formant whose COG matched that of the formant that had been removed. The frequency separation between the two component resonances was then systematically varied between 1.5 and 3.5 barks and a singleinterval 2AFC vowel identification task was used to obtain estimates of vowel quality for each series step. Results will be discussed in terms of whether the spectral integration effects within the 3.5 bark decline as the frequency separation between the resonance components increases. [Work supported by NIDCD R01DC00679-01A1.]

An Exploration of Skin Acceleration Level as a Measure of Phonatory Function in Singing

Two kinds of fluctuations are observed in phonetogram recordings of singing. Sound pressure level (SPL) can vary due to vibrato and also due to the effect of open and closed vowels. Since vowel variation is mostly a consequence of vocal tract modification and is not directly related to phonatory function, it could be helpful to suppress such variation when studying phonation. Skin acceleration level (SAL), measured at the jugular notch and on the sternum, might be less influenced by effects of the vocal tract. It is explored in this study as an alternative measure to SPL. Five female singers sang vowel series on selected pitches and in different tasks. Recorded data were used to investigate two null hypotheses: (1) SPL and SAL are equally influenced by vowel variation and (2) SPL and SAL are equally correlated to subglottal pressure (P(S)). Interestingly, the vowel variation effect was small in both SPL and SAL. Furthermore, in comparison to SPL, SAL correlated weakly to P(S). SAL exhibited practically no dependence on fundamental frequency, rather, its major determinant was the musical dynamic. This results in a non-sloping, square-like phonetogram contour. These outcomes show that SAL potentially can facilitate phonetographic analysis of the singing voice.

Auditory and phonetic perception of vowels in children with apraxic speech disorders

The aim of this study was to assess auditory and phonetic perceptual processing of vowels in children with apraxic disorders, who demonstrated clinically with only a speech output deficit. Two experiments were conducted. In the preparatory Experiment 1 series of vowels were constructed by moving formant frequencies away from the extreme values in the vowel space in the direction of a ‘neutral‐vowel position’. These were presented to adults and children with no speech‐language involvement. Based on identification performance low‐redundancy vowels were selected, which served as the end‐points of two vowel continua: /i/‐/I/ and /a/‐/α/. In Experiment 2 these continua were used in identification and discrimination tasks, presented to 11 children with apraxic speech problems (aged 6:11 to 9:6 years) and 12 normally developing children. The results showed poorer perception of vowels for the children with apraxic speech problems than for the control children for both continua. Identification functions indicated poorer phonetic processing; discrimination functions indicated poorer auditory processing. Furthermore, a combination of perception measures (identification and discrimination) proved to have a high differential and clinical value for the assessment of children with apraxic speech problems. The results support the view that subtle (subclinical) auditory processing deficits make part of speech output disorders.

Somali (review)

Reviewed by: Somali by John Saeed Gonzalo Rubio Somali. By John Saeed. (London Oriental and African language library 10.) Amsterdam & Philadelphia: John Benjamins, 1999. Pp. xv, 295. This is an outstanding overview of Somali, one of the most widely spoken Cushitic languages, with detailed linguistic descriptions of its general framework (1–6), phonology (7–51), morphology and word formation (53–162), syntax (163–228), and pragmatics of discourse (229–250). Two appendixes list all the data sources (251–52) and offer some parsed sample texts (253–65). The volume ends with a bibliography (279–87) and indexes (289–95). It is somewhat annoying that the editors opted for final notes (267–77) instead of the more useful footnotes. Whereas such a choice might be somehow justified in works of haute divulgation, it is quite objectionable in the case of scholarly publications. The presentation of the materials and the descriptive analysis used here are impeccable and make this work exceedingly practical for linguists and students working in Somali, Cushitic, or Afroasiatic studies. The overall methodology employed seems to depend on the typologico-functional approach, as formulated by Talmy Givón and others. This approach has proven especially fertile in the description and close analysis of diverse languages of the world as opposed to other perhaps more heavily theoretical but rather English-centric schools. The description of the vocalic system (11–16) argues in favor of two series of vowels differentiated by the ATR feature: back [−ATR] and front [+ATR] vowels. Saeed is in agreement with the vast majority of scholars of Somali in his adherence to B. W. Andrzejewski’s description of a dual system opposing fronting vs. retracting for the Somali vowels; see also Annarita Puglielli, ‘Somali phonology’, (Phonologies of Asia and Africa, 523–25. Winona Lake: Eisenbrauns, 1997)—this is not included in S’s bibliography. As I have mentioned elsewhere (Language 75.141), one still misses a discussion—or a simple mention—of Joseph Pia’s proposal concerning a multiply-tiered system (Current progress in Afro- Asiatic linguistics, Amsterdam: John Benjamins, 1978, 463–75). This well-written and intelligibly explained work will help to improve the knowledge of Somali among linguists and readers interested in Somali matters and in the peoples and the languages of the Horn of Africa. S should be thanked and congratulated for this delightful book. Gonzalo Rubio Ohio State University Copyright © 2001 Linguistic Society of America

Perception of Vowel-to-Vowel Transitions with Different Formant Trajectories

In this paper, the perceptual effects of vowel-to-vowel transitions determined by different temporal variations of model parameters which specify the shapes of the vocal tract area function are investigated. It is shown that, (a) the method of deformation of the vocal tract area function between two targets can be perceptually important and (b) conversely, within certain limits, the time course of parameters from one state to another, and the precise synchronization of two parameters is not important for the correct identification of a vowel series. These characteristics are necessary but not sufficient to prove the existence of a phonetic gesture percept.

Neighboring spectral content influences vowel identification.

Four experiments explored the relative contributions of spectral content and phonetic labeling in effects of context on vowel perception. Two 10-step series of CVC syllables ([bVb] and [dVd]) varying acoustically in F2 midpoint frequency and varying perceptually in vowel height from [delta] to [epsilon] were synthesized. In a forced-choice identification task, listeners more often labeled vowels as [delta] in [dVd] context than in [bVb] context. To examine whether spectral content predicts this effect, nonspeech-speech hybrid series were created by appending 70-ms sine-wave glides following the trajectory of CVC F2's to 60-ms members of a steady-state vowel series varying in F2 frequency. In addition, a second hybrid series was created by appending constant-frequency sine-wave tones equivalent in frequency to CVC F2 onset/offset frequencies. Vowels flanked by frequency-modulated glides or steady-state tones modeling [dVd] were more often labeled as [delta] than were the same vowels surrounded by nonspeech modeling [bVb]. These results suggest that spectral content is important in understanding vowel context effects. A final experiment tested whether spectral content can modulate vowel perception when phonetic labeling remains intact. Voiceless consonants, with lower-amplitude more-diffuse spectra, were found to exert less of an influence on vowel perception than do their voiced counterparts. The data are discussed in terms of a general perceptual account of context effects in speech perception.

Formant Pattern Ambiguity of Vowel Sounds

The formant frequencies of a particular vowel vary according to the speaker group and to Coarticulatoin. Therefore, overlapping formant patterns of different vowels are commonly related to sex and age differences and to coarticulation, and are considered to concern mainly the F1-F2 pattern of adjacent vowels. However, several studies have reported indications of a correlation between the lower formant frequencies and F0. as well as of the appearance of different formant numbers relevant to vowel identity. As a consequence, the overlap between formant patterns of different vowels might be more substantial than has traditionally been assumed. Within the present study, therefore, the extent to which a given formant pattern can represent different vowels was investigated for natural Swiss German vowels produced monotonously and in isolation by men, women and children at F0 of 85–870 Hz. Similar formant patterns were found for vocalizations of different vowels with both small and large phonetic distances, and within the entire frequency ranges of the formants relevant for phoneme identity. For vowel sounds displaying ambiguous formant patterns, the main spectral characteristics related to differences in their perceptual identity were found to concern F0 and relative formant amplitudes. Results are given in exemplary vowel series, and consequences for the psychophysics of the vowel are discussed.

Speaking rate effects on vowel identification in natural sentence contexts

Previous research [T. L. Gottfried, J. L. Miller, and P. E. Payton, Phonetica 47, 155–172 (1990)] has indicated for synthetic speech that vowel identification changes according to the speaking rate of the sentence context, even when the vowel contrast is specified experimentally only by spectral differences. This study tests listeners’ identification of /i/ vs /I / in stimulus continua created by altering formant values and durations of the target syllables ‘‘beat’’ and ‘‘bit’’ in naturally produced sentences. These targets were spoken in two sentence frames, one spoken at a moderate rate of speech and the other compressed to twice the speaking rate. Longer syllable durations significantly shifted the boundary of the spectral continuum, so that there were more ‘‘beat’’ responses as compared to shorter durations. However, unlike what was found for synthetic vowel series, there was no significant effect of sentence rate on the number of ‘‘beat’’ responses. Reasons for this different outcome, and the relevance of these findings for theories of rate normalization, are discussed. [Work supported by Lawrence University Faculty Grant.]

Isolating the critical segment of AE /r/ and /l/ to enhance non-native perception

Brief synthetic stop-consonant+vowel syllables as short as 20 ms are reliably identified for place of articulation. Further, removal of the steady-state vowel in synthetic stop-consonant+vowel series improves discrimination performance, but does not affect identification. Subsequently, vowel truncation of consonant-vowel syllables has been used to improve native English speaker’s perception of Hindi dental and retroflex stop-consonants. Whether such truncation techniques are effective for native speakers of Japanese on the AE liquids /r/ and /l/ was examined. The /r/ or /l/ segment from 50 minimal word pairs, produced in five phonemic positions, was isolated by truncating the surrounding phonemic environment. Six durations of the segment were created: 60, 100, 140, 180, 220 ms, and full-length (unedited productions). Native Japanese speakers were asked to identify the target phoneme for each segment, which was presented in a random sequence but blocked by speaker and segment duration. Segment duration significantly affected performance. Interestingly, while longer segments were perceived better than full-length stimuli, shorter segments were not perceived as well. Identification training with these stimuli confirmed the effectiveness of such manipulations. The implications of these findings are discussed with regard to theories of cross-language speech perception and methods of training non-native contrasts.

An acoustic study of Italian unstressed mid-vowels

In Italian, the contrast between the two mid front vowels |eε| and between the two mid back vowels |o■| is neutralized in unstressed syllables, where the two pairs merge into the unstressed mid ‘‘archiphonemes’’ |EO|. In this study, an instrumental investigation of the quality of the unstressed mid-vowels is presented to see if neutralization is complete (phonetically), and to determine the nature of |EO| in different contexts. The preliminary results show that there is no statistical difference between an |E| alternating with |e|, an |E| alternating with |ε|, and an |E| that does not alternate with any stressed vowel. The same holds for the mid-back vowel series. In addition, pre-tonic |EO| are realized as mid-high vowels (their average F1 values indicate that they are slightly higher than stressed |eo|), while post-tonic |EO| are realized as mid-low vowels (their average F1 values indicate that they are very similar to stressed |ε■|). In neither case is a significant centralization effect observed. Finer grain distinctions between pre and post-tonic |EO| in different contexts will be examined further.

ON ‘TENSENESS’ IN MODERN ICELANDIC1

ABSTRACTModern Icelandic presents phonological oppositions between two series of stop consonants and two series of vowels. Unlike in Old Icelandic, these are not phonetically characterized by voice in the former case or length in the latter, and the modern distinguishing features have commonly been described in both cases as ‘tenseness’ (versus ‘laxness’). There has, however, been no adequate definition of this feature, and the distinctions involved are fully describable in terms of various more traditional and observable processes. There is in fact no one feature common to the consonantal and vocalic oppositions in question; and even within each of these some problems arise when one attempts to find a single defining criterion. Various possible solutions are examined and evaluated. Some consideration is also given to the relevant historical and dialectal dimensions.

The role of multiple acoustic properties in specifying the internal structure of phonetic categories

The extent to which a phonetic segment may be specified by more than one acoustic property was investigated using the trading relation between vowel duration and closure duration, two properties known to play a role in specifying the voicing contrast for intervocalic bilabial stops. Specifically asked was whether a change in preceding vowel duration results in a comprehensive remapping between closure duration and phonetic category, or whether trading effects are confined to the category boundary. Two series of disyllables were created ranging from /aba/ through /apa/ to * /apa/ (an exaggerated /p/) having initial vowel durations of 153 and 250 ms, respectively. Closure duration in each series varied from 20 to 400 ms. A preliminary experiment revealed a standard trading relation, in that the /b/‐/p/category boundary was located at a longer closure duration for the stimuli with a long, compared to a short, preceding vowel. In the main experiment, listeners were asked to judge each disyllable in each series for the goodness of its consonant as a member of the /p/ category. For both series the /p/ category was perceived as having internal structure, with a limited range of stimuli being judged as the best exemplars. Furthermore, the range of best exemplars for the long vowel series was displaced relative to that for the short vowel series toward longer values of closure duration. These findings indicate that the acoustic properties in question trade against each other not only at the phonetic category boundary but also within the category. This results in a comprehensive remapping of phonetic category structure similar to that observed in earlier research for changes in speaking rate. [Work supported by NIH.]