F2 Onset Frequency Research Articles

Influence of /r/ on burst spectra for stop place

Bilabial and lingua-alveolar stops in 20 minimally contrastive word pairs in both singleton and stop /r/ (cluster) contexts were recorded from two girls, one with facial paralysis (CFP) and one with normal facial movement (CNM). Auditory identification of these productions by 12 listeners revealed a significant place by context interaction for CFP. Identification scores were high for her lingua-alveolar stops (99.3% singletons; 94.8% clusters). Identification scores for her bilabial stops were lower, with singletons being significantly lower (57.8%) than clusters (77.8%). Acoustic cues for stop place (F2 onset frequency, VOT, mean and skewness of burst spectra) were measured for all word productions. For both girls, F2 onset and VOT measures were lower for bilabial than lingua alveolar stops in singletons and clusters. CFP’s burst spectra for bilabials had a higher mean and more negative skewness than lingua alveolars in singletons and clusters. CNM’s burst spectra for bilabials had a lower mean and more positive skewness than lingua alveolars in singletons (expected) but a higher mean and more negative skewness for bilabials than lingua alveolars in clusters (unexpected). This unexpected finding may account for CFP’s higher identification rates for bilabial clusters (all acoustic cues ‘‘fit’’ those of a child with normal facial movement).

Ability in perceiving nonnative contrasts: performance on natural and synthetic speech stimuli.

The perception of the distinction between /r/ and /l/ by native speakers of American English and of Japanese was studied using natural and synthetic speech. The American subjects were all nearly perfect at recognizing the natural speech sounds, whereas there was substantial variation among the Japanese subjects in their accuracy of recognizing /r/ and /l/ except in syllable-final position. A logit model, which additively combined the acoustic information conveyed by F1-transition duration and by F3-onset frequency, provided a good fit to the perception of synthetic /r/ and /l/ by the American subjects. There was substantial variation among the Japanese subjects in whether the F1 and F3 cues had a significant effect on their classifications of the synthetic speech. This variation was related to variation in accuracy of recognizing natural /r/ and /l/, such that greater use of both the F1 cue and the F3 cue in classifying the synthetic speech sounds was positively related to accuracy in recognizing the natural sounds. However, multiple regression showed that use of the F1 cue did not account for significant variance in natural speech performance beyond that accounted for by the F3 cue, indicating that the F3 cue is more important than the F1 cue for Japanese speakers learning English. The relation between performance on natural and synthetic speech also provides external validation of the logit model by showing that it predicts performance outside of the domain of data to which it was fit.

Place of articulation and first formant transition pattern both affect perception of voicing in English

Voice onset time (VOT) and first-formant (F1) transition characteristics are both important phonetic cues for voicing categorization of pretonic stops in English. These cues also covary with place of articulation in production, such that bilabials are produced with shorter VOT values, and velars are produced with longer VOT values. This study investigated the effects of place of articulation and F1transition pattern on voicing categorization of VOT continua. Listeners categorized for voicing and place of articulation the stop consonant in VCV and CV stimuli for which the variables F1transition pattern, place of articulation (bilabial, alveolar, and velar), and VOT were fully crossed. A logistic regression analysis shows that bilabial and alveolar stimuli were more likely than velar stimuli to be classified as voiceless, largely independent of F1transition pattern. Increasing F1onset frequency and shortening transition duration also made voiceless judgments more likely. The magnitude of theF1 transition effects was considerably larger than the place of articulation effect. The results are inconsistent with models of consonant classification in which acoustic–phonetic cues for place of articulation are not involved in the perception of the voicing contrast. However, the observed perceptual interaction between place of articulation and voicing may be consistent with either a feature- or segment-based model of consonant classification.

An acoustic study of coarticulation in dysarthric speakers with Parkinson disease.

Anticipatory vowel coarticulation in obstruent + vowel syllables was studied for 9 dysarthric males with mild to moderate idiopathic Parkinson disease (PD) and a group of 10 healthy males. In addition to studying coarticulatory differences for speakers' habitual rate repetitions, a graded speaking rate task was used to investigate the effect of rate variation on coarticulation. The ratio of F2 onset frequency/F2 target frequency was used to infer coarticulation (i.e., ratios of 1.0 indicate complete vowel harmony at vowel onset, and ratios greater than 1.0 indicate relatively less coarticulation). Because F2 onset frequency is a relatively novel measure of coarticulation, a more common measure of anticipatory vowel coarticulation-fricative F2-was obtained for a subset of the speech stimuli and compared to F2 onset measures. Ratios of fricative F2/F2 target frequency also were computed to infer coarticulation and the results were compared to the F2 onset/F2 target ratios. The results indicated a modest relationship between F2 onset frequency and fricative F2 for both speaking tasks and speaker groups, with a stronger relationship for the graded rate task. Regression analyses for the graded rate task further indicated that longer vowel durations were associated with larger ratios for both speaker groups. Thus, coarticulation tended to increase with faster rates and decrease with slower rates, although the relationship was more systematic for control speakers. Ratios for the habitual task tended to be smaller for speakers with PD, suggesting increased coarticulation relative to control speakers. This effect was not entirely attributable to the more rapid speaking rates for speakers with PD. Because habitual rate F2 onset/F2 target ratios tended to underestimate the extent to which coarticulation was increased for speakers with PD-compared to the fricative F2/F2 target ratios-measures other than F2 onset should be used to infer absolute differences in extent of coarticulation for speakers with PD and healthy controls.

Auditory-visual context effects on the perception of voicing

Perception of an auditory continuum can be changed by a preceding visual context [K. Green and L. Norrix, in press, JEP: HPP]. This study examines if a visual context for /l/, as in /abla/, influences the voicing distinction for an auditory /aba/-/apa/ continuum. Voice onset time (VOT) and F1 onset frequency were measured during production of /aba/, /apa/ (singleton conditions) and /abla/, /apla/ (cluster conditions). Findings indicate longer VOTs in /apla/ compared to /apa/ and lower F1 onset frequencies for the cluster compared to singleton conditions. An auditory-only (AO) continuum (/aba/–/apa/) was created and presented to listeners for identification. In an auditory-visual (AV) condition each auditory token was paired with a face saying /abla/ and presented for identification. If subjects use knowledge about coarticulation and its effects on VOT during production, then they should make more /b/ responses in the AV compared to AO condition. In contrast, if they recover F1 information from the visual articulation, then more /p/ responses should be obtained. Findings indicate more /p/ responses in the AV compared to AO condition, suggesting that perceivers are recovering information from visual articulations (perhaps tongue height) rather than using stored knowledge about coarticulatory effects on VOT. [Work supported by NSF.]

Neural representation of consciously imperceptible speech sound differences.

The concept of subliminal perception has been a subject of interest and controversy for decades. Of interest in the present investigation was whether a neurophysiologic index of stimulus change could be elicited to speech sound contrasts that were consciously indiscriminable. The stimuli were chosen on the basis of each individual subject's discrimination threshold. The speech stimuli (which varied along an F3 onset frequency continuum from /da/ to /ga/) were synthesized so that the acoustical properties of the stimuli could be tightly controlled. Subthreshold and suprathreshold stimuli were chosen on the basis of behavioral ability demonstrated during psychophysical testing. A significant neural representation of stimulus change, reflected by the mismatch negativity response, was obtained in all but 1 subject in response to subthreshold stimuli. Grand average responses differed significantly from responses obtained in a control condition consisting of physiologic responses elicited by physically identical stimuli. Furthermore, responses to suprathreshold stimuli (close to threshold) did not differ significantly from subthreshold responses with respect to latency, amplitude, or area. These results suggest that neural representation of consciously imperceptible stimulus differences occurs and that this representation occurs at a preattentive level.

Acoustic characteristics of English fricatives

This study constitutes a large-scale comparative analysis of acoustic cues for classification of place of articulation in fricatives. To date, no single metric has been found to classify fricative place of articulation with a high degree of accuracy. This study presents spectral, amplitudinal, and temporal measurements that involve both static properties (spectral peak location, spectral moments, noise duration, normalized amplitude, and F2 onset frequency) and dynamic properties (relative amplitude and locus equations). While all cues (except locus equations) consistently serve to distinguish sibilant from nonsibilant fricatives, the present results indicate that spectral peak location, spectral moments, and both normalized and relative amplitude serve to distinguish all four places of fricative articulation. These findings suggest that these static and dynamic acoustic properties can provide robust and unique information about all four places of articulation, despite variation in speaker, vowel context, and voicing.

Stress-induced variation in F2 trajectories as evidence for coproduction in CV syllables.

This study explored the extent to which a model of the acoustic consequences of overlapping, sliding consonantal and vocalic gestures was used to account for stress-induced changes in F2 trajectories occurring in test words embedded in a carrier phrase. Three stress conditions were studied including contrastive stress on test words (CS), contrastive stress on the content word preceding test words (U-CS), and non contrastive stress on test words (NS). F2 onset frequency was used to quantify the extent to which adjacent consonantal and vocalic gestures in stop consonant + vowel syllables were coproduced (i.e., overlapped) in the different stress conditions. The predicted relationship between F2 onset frequency and temporal variation in trajectories was also examined within and across stress conditions. In addition, the effects of stress-induced variation in articulatory scaling on F2 onset frequency were studied and factored into the interpretation of the results. The results indicated that F2 onset frequencies tended to differ for stress conditions characterized by large differences in prominence. Regression analyses predicting temporal variation in trajectories from F2 onset frequency accounted for part of the variance within and across stress conditions. Taken together, the results suggested that a model of overlapping, sliding gestures accounts for only some of the stress-induced variability in F2 trajectories in the current study.

Effects of first formant onset properties on voicing judgments of prevocalic stops without F1 cutback

This study examines the effects of first formant (F1) onset properties on voicing judgments of prevocalic stop consonants in contexts that do not exhibit the F1 cutback covariation present in pretonic and utterance-initial voicing contrasts. Previous research [K. R. Kluender and A. J. Lotto, J. Acoust. Soc. Am. 95, 1044–1052 (1994)] strongly suggests that F1 onset frequency effects on voicing judgements are due to auditory factors, and not to listener experience with natural covariance with F1 cutback and F1 onset frequency. The present study extends the previous findings to voicing contrasts in contexts that are not signaled by F1 cutback, such as before unstressed vowels in English. Categorization data were collected for a continuum between rabid (/b/) and rapid (/p/) in which both closure duration and F1 onset properties were manipulated using LPC resynthesis of natural speech. Preliminary results indicate that lower F1 onset frequencies of the unstressed vowel condition more voiced percepts of the preceding stop consonant, consistent with findings for utterance-initial stop consonants.

Place of articulation for consonants: Comparing nasals and stops, and syllable position

This study reports on measurements of transitions of the second formant (F2) for syllable-initial and syllable-final nasal and stop consonants in English, as cues to place of articulation. F2 values and the amount and direction of F2 changes in a 20 ms interval were determined in vowels adjacent to consonant ‘‘implosion’’ and release. The corpus included consonants in sentences and isolated nonsense syllables. The F2 transitions for a given place of articulation are roughly similar for nasals and stops for different syllable positions, as expected, but some interesting differences exist. Since the stop-consonant bursts influence the spectrum sampling point relative to release, there are some shifts in F2 onset frequencies for stops, as compared to nasals. Syllable-initial alveolars show additional manner differences, which can be attributed to differences in tongue configurations for alveolar nasals relative to stops in back-vowel contexts. These latter shifts are tentatively ascribed to differential coarticulatory effects for /n/ versus /d/ and /t/ differences which might be related to the lack of a distinctive syllable-initial velar nasal in English. There are also systematic shifts in F2 transitions as a function of syllable position, for both nasals and stops. [Supported in part by NIH Grants DC00075 and DC02525.]

Can a model of overlapping gestures account for scanning speech patterns?

A simple acoustic model of overlapping, sliding gestures was used to evaluate whether coproduction was reduced for neurologic speakers with scanning speech patterns. F2 onset frequency was used as an acoustic measure of coproduction or gesture overlap. The effects of speaking rate (habitual versus fast) and utterance position (initial versus medial) on F2 frequency, and presumably gesture overlap, were examined. Regression analyses also were used to evaluate the extent to which across-repetition temporal variability in F2 trajectories could be explained as variation in coproduction for consonants and vowels. The lower F2 onset frequencies for disordered speakers suggested that gesture overlap was reduced for neurologic individuals with scanning speech. Speaking rate change did not influence F2 onset frequencies, and presumably gesture overlap, for healthy or disordered speakers. F2 onset frequency differences for utterance-initial and -medial repetitions were interpreted to suggest reduced coproduction for the utterance-initial position. The utterance-position effects on F2 onset frequency, however, likely were complicated by position-related differences in articulatory scaling. The results of the regression analysis indicated that gesture sliding accounts, in part, for temporal variability in F2 trajectories. Taken together, the results of this study provide support for the idea that speech production theory for healthy talkers helps to account for disordered speech production.

Speaking-rate-induced variability in F2 trajectories.

This study examined speaking-rate-induced spectral and temporal variability of F2 formant trajectories for target words produced in a carrier phrase at speaking rates ranging from fast to slow. F2 onset frequency measured at the first glottal pulse following the stop consonant release in target words was used to quantify the extent to which adjacent consonantal and vocalic gestures overlapped; F2 target frequency was operationally defined as the first occurrence of a frequency minimum or maximum following F2 onset frequency. Regression analyses indicated 70% of functions relating F2 onset and vowel duration were statistically significant. The strength of the effect was variable, however, and the direction of significant functions often differed from that predicted by a simple model of overlapping, sliding gestures. Results of a partial correlation analysis examining interrelationships among F2 onset, F2 target frequency, and vowel duration across the speaking rate range indicated that covariation of F2 target with vowel duration may obscure the relationship between F2 onset and vowel duration across rate. The results further suggested that a sliding based model of acoustic variability associated with speaking rate change only partially accounts for the present data, and that such a view accounts for some speakers' data better than others.

Spectral contrast in perception of VCV syllables

The ability of listeners to recover speech information, despite dramatic articulatory and acoustic assimilation between phonetic units, is central to understanding perception of fluent speech. The present studies investigate the effect of preceding acoustic stimuli on identification of a series of synthetic CVs varying acoustically in F2 onset frequency and perceptually from /ba–da/. Results reveal that naturally produced /i/ and /u/ modulate identification responses to the CV series independent of speaker, suggesting that the effect of preceding V on CV identification is not critically dependent on precise matching of articulatory characteristics between V and CV. Moreover, nonspeech FM glides modeling F2 transitions of natural Vs shift CV identification in a similar manner, suggesting that quite general auditory mechanisms may contribute to observed shifts in CV identification. Further experiments reveal the effect of preceding V maintains for dichotic presentation and persists, but lessens, with intersyllabic silent intervals increasing to 200 ms. All results can be recast as cases of spectral contrast whereby spectral characteristics of preceding sound modulate perception of CV. These data suggest general auditory processes may play a substantive role in perceptual accommodation of coarticulated speech and provide clues to the nature of the putative auditory mechanisms. [Work suported by NSF.]

General contrast effects in speech perception: effect of preceding liquid on stop consonant identification.

When members of a series of synthesized stop consonants varying acoustically in F3 characteristics and varying perceptually from /da/ to /ga/ are preceded by /al/, subjects report hearing more /ga/ syllables relative to when each member is preceded by /ar/ (Mann, 1980). It has been suggested that this result demonstrates the existence of a mechanism that compensates for coarticulation via tacit knowledge of articulatory dynamics and constraints, or through perceptual recovery of vocal-tract dynamics. The present study was designed to assess the degree to which these perceptual effects are specific to qualities of human articulatory sources. In three experiments, series of consonant-vowel (CV) stimuli varying in F3-onset frequency (/da/-/ga/) were preceded by speech versions or nonspeech analogues of /al/ and /ar/. The effect of liquid identity on stop consonant labeling remained when the preceding VC was produced by a female speaker and the CV syllable was modeled after a male speaker's productions. Labeling boundaries also shifted when the CV was preceded by a sine wave glide modeled after F3 characteristics of /al/ and /ar/. Identifications shifted even when the preceding sine wave was of constant frequency equal to the offset frequency of F3 from a natural production. These results suggest an explanation in terms of general auditory processes as opposed to recovery of or knowledge of specific articulatory dynamics.

Speaking rate-induced acoustic variability in Parkinson’s disease

The idea that speech consists of overlapping gestures that slide with respect to one another suggests that measurable kinematic, and hence, acoustic differences should characterize utterances produced with varying degrees of gesture overlap. [Saltzman and Munhall, Ecolog. Psych. 1, 333–382 (1989)]. Further research suggests that a simple acoustic model of overlapping gestures [Weismer et al., J. Phon. 23, 149–164 (1995)] accounts for some of the spectral and temporal variability in speech produced by individuals with apraxia of speech. One implication is that articulatory deficits associated with motor speech disorders can be modeled, in part, as disruptions in gesture overlap. The extent to which a model of overlapping gestures accounts for acoustic variability in other motor speech disorders is largely unknown. The present study explored the extent to which a simple acoustic model of overlapping gestures accounted for speaking-rate induced variability in F2 trajectories produced in target words by individuals with Parkinson’s disease (PD), young adults, and older adults. F2 onset frequency was used as an acoustic measure of overlapping consonantal and vocalic gestures; F2 onset frequency also was used to predict across-repetition temporal variability in F2 trajectories.

Effects of place of articulation and first formant onset properties on voicing judgments

The categorization of voiced and voiceless syllable-initial stops is dependent primarily on voice onset time (VOT), but has been proposed to be sensitive to first formant (F1) transitions and place of articulation. The present study tested the effect of both F1 transition and place cues on the categorization of voiced and voiceless stops in a /■Ca/ context (second syllable stress). Three different stop properties were independently manipulated: VOT, F1 onset frequency, and place of articulation. VOT was varied along a 0 to 65-ms continuum. Four different F1 onset frequencies for the /Ca/ transition were used. Place (labial, alveolar, and velar) was cued by both the release burst and higher formant transitions into and out of the stop. The effects of F1 transition and place on voicing judgments were measured by shifts in the VOT identification boundary. The effect of F1 on voicing judgments would be interpreted as resulting from a general mechanism of audition nonspecific to speech, following Kluender [J. Acoust. Soc. Am. 90, 83–96 (1991)] who found that categorization of burstless CV stimuli from a VOT continuum is sensitive to F1 onset frequency, but not place. In contrast, an effect of place on voicing judgments would be interpreted as the result of a phonological category determining phonetic criteria, in this case place determining VOT thresholds.

Acoustic correlates of segmentalized speech patterns in motor speech disorders

It has been argued that principles of gestural-based speech production theory may account for speech production aberrancies in motor speech disorders (MSDs) [Weismer et al., J. Phon. 23, 111–126 (1995)]. In the present study, four individuals with MSDs exhibiting segmentalized speech patterns (e.g., perceptual impression of speech being ‘‘metered out’’ segment by segment) and four controls produced CV words in carrier phrases at fast and habitual rates. Degree of gesture overlap was indexed by F2 onset frequency. Descriptive measures ofF 2 onset were used to evaluate whether segmentalized speech patterns in MSDs were associated with reduced gesture overlap in comparison to controls. Temporal variability in F2 formant trajectories for all speakers was modeled as the mutual sliding of adjacent gestures using regression analyses. Three disordered speakers showed evidence of reduced overlap for targets produced at habitual rates in comparison to controls; this difference generally was maintained at fast rates. For all speakers, fast rates were associated with reductions in both gesture magnitude and overlap relative to habitual rates. Regression analyses suggested that temporal variability in F2 trajectories of select targets produced at habitual rates may be modeled as the mutual sliding of gestures. Individual differences and stimuli effects are discussed.

Internal structure of phonetic categories: Evidence for within-category trading relations

Phonetically relevant acoustic properties perceptually trade against each other at phonetic category boundaries. The present investigation used a category goodness judgment task to examine whether such properties also trade against each other in specifying the best exemplars of phonetic categories. The experiments focused on the say–stay distinction, specified by F1 onset frequency and silence duration preceding F1 onset. The main experiment demonstrated a within-category trading relation, such that as F1 onset frequency increased from 230 to 430 Hz a longer silence was required for stimuli to be judged the best exemplars of stay. Two follow-up experiments explored the robustness of this effect. Taken together, the findings underscore the importance of multiple acoustic properties in specifying the internal structure of phonetic categories.

Acoustic elements of speechlike stimuli are reflected in surface recorded responses over the guinea pig temporal lobe

Auditory evoked potentials measured from the guinea pig temporal lobe surface reflect acoustic elements of synthesized speech syllables. Eliciting stimuli included a four formant anchor stimulus /ba/, with a 40-ms formant transition duration. The other stimuli differed from /ba/ along simple acoustic dimensions. The /pa/ stimuli differed on a VOT continuum; /da/ stimuli had a higher frequency F2 onset; /wa/ had a longer (80 ms) formant transition duration; and /bi/ differed in three vowel formant frequencies. The /ba/ and /da/ onset response latencies decreased systematically with increasing F2 onset frequency. The response to the /pa/ voicing increased in latency with increasing VOT and showed a physiologic discontinuity at VOT of 15-20 ms. Responses to /ba/ and /wa/ showed similar onset morphology but significant amplitude differences at latencies corresponding to vowel onset. Significant amplitude differences in /ba/ and /bi/ responses corresponded in latency to both consonant and vowel portions of the syllables. Similar to previous reports in the awake monkey for VOT, these results demonstrate in the anesthetized guinea pig that acoustic elements essential to speech perception are reflected in aggregate response of ensembles of cortical neurons.

An acoustic evaluation of variation in the overlap of consonant and vowel gestures induced by speaking rate change.

There is a growing body of empirical data supporting a gestural‐based view of speech production in which adjacent speech gestures temporally and spatially overlap one another. Recent work suggests that certain acoustic measures are sensitive to changes in the overlap of adjacent consonant and vowel gestures [G. Weismer et al., J. Phon. 23, 111–126 (1995); E. Zsiga, J. Phon. 22, 121–140 (1994)]. Utilizing a graded speaking rate task, eight speakers produced 160 repetitions of ten target words embedded in a carrier phrase. F2 onset frequency, measured at the consonant–vowel boundary of test syllables, was used to index the degree of spectral‐temporal overlap of adjacent consonant and vowel gestures. Regression analyses were used to evaluate the extent to which F2 onset predicted temporal variability in F2 formant trajectories due to speaking rate change. Results suggest that consonant and vowel gestures do not simply temporally slide away from each other with slowed speaking rate. Rather, modifications in the form and magnitude of vowel gestures are required to account for the present acoustic data. An empirically based acoustic model of variation in gesture overlap induced by speaking rate change is offered. Individual speaker differences also are discussed.