F2 Onset Research Articles

Is fundamental frequency a cue to aspiration in initial stops?

One production and one perception experiment were conducted to investigate the interaction of consonant voicing and fundamental frequency at the onset of voicing (onset f0) in Cantonese, a tonal language. Consonantal voicing in English can affect onset f0 up to 100 ms after voicing onset, but existing research provides inconclusive information regarding the effects of voicing on f0 in tonal languages where f0 variability is constrained by the demands of the lexical tone system. Previous research on consonantal effects on onset f0 provides two contrasting theories: These effects may be automatic, resulting from physiological constraints inherent to the speech production mechanism or they may be controlled, produced as part of a process of cue enhancement for the perception of laryngeal contrasts. Results of experiment 1 showed that consonant aspiration affects onset f0 in Cantonese only within the first 10 ms following voicing onset, comparable to results for other tonal languages. Experiment 2 showed that Cantonese listeners can use differences in onset f0 to cue perception of the voicing contrast, but the minimum extent of f0 perturbation necessary for this is greater than is found in Cantonese production, and comparable to that observed in acoustic studies of nontonal languages. These results suggest that consonantal effects on onset f0 are at least partially controlled by talkers, but that their role in the perception of voicing/aspiration may be a consequence of language independent properties of audition rather than listeners' experience with the phonological contrasts of a specific language.

Selective attention and perceptual learning of speech

Phonetic experience can change the perceptual distance between speech sounds, increasing both within-category similarity and between-category distinctiveness. Such warping of perceptual space is frequently characterized in terms of changes in selective attention: Listeners are assumed to attend more strongly to category-differentiating features while ignoring less relevant ones. However, the link between the distribution of selective attention and categorization-related differences in perceptual distance has not been empirically demonstrated. To explore this relationship, listeners were given 6 h of training to categorize sounds according to one of two acoustic features while ignoring the other. The features were voice onset time and onset f0, which are normally correlated and can both serve as a cue to consonant voicing. Before and after training, listener’s performance on a Garner selective attention task was compared with assessment of the perceptual distance between tokens. Results suggest that training can induce both warping of perceptual space and changes in the distribution of selective attention, but the two phenomena are not necessarily related. Results are consistent with a two-stage model of perceptual learning, involving both preattentive adjustment of acoustic cue weighting and higher-level changes in the distribution of selective attention between acoustic cues. [Work supported by NIH-NIDCD 1R03DC006811.]

The effect of stress on consonant vowel (CV) coarticulation: Decoupling of the CV bond

Two common sources of variation in the speech signal arise from (1) naturally imposed speaking conditions (e.g., stress, tempo), overlaid onto (2) inherent contextual influences. Coarticulatory analyses are affected by both sources of variation, but rarely are they dissociated. The focus of this study was to derive quantitative methods to assess the effect of emphatic vowel stress on preceding stop consonant onsets, apart from, and independent of, anticipatory effects of changing vowel contexts. Three speakers produced V1CV2 tokens, with emphatic stress either on V1 or V2; C=/bdg/. There were six V1 and 10 V2 vowel contexts. Novel methods were applied to both standard locus equation scatterplots and multiple regression analyses to isolate F2 midvowel (Hz) and F2 onset (Hz) adjustments as a function of emphatic stress. The analyses show how stress differentially affects the V-midpoints and C-onsets for labial /b/ relative to lingual stops /dg/.

Sequential contrast or compensation for coarticulation?

English listeners identify a stop ambiguous between /t/ and /k/ more often as ‘‘k’’ after /s/ than /sh/ (Mann and Repp, 1981). Judgments shift similarly after a fricative ambiguous between /s/ and /sh/ when its identity is predictable from a transitional probability bias but perhaps not from a lexical bias (Pitt and McQueen, 1998; cf. Samuel and Pitt, 2003). In replicating these experiments, we add a discrimination task to distinguish between the predictions of competing explanations for these findings: listeners respond ‘‘k’’ more often after /s/ because they compensate for the fronting of the stop expected from coarticulation with /s/ or because a stop with an F3 onset frequency midway between /t/’s high value and /k/’s low value sounds lower after the /s/’s high-frequency energy concentration. The second explanation predicts listeners will discriminate /s-k/ and /sh-t/ sequences better than /s-t/ and /sh-k/ sequences because sequential contrast exaggerates the spectral differences between /s-k/’s high-low intervals and /sh-t/’s low-high intervals and distinguishes them more perceptually than /s-t/’s high-high intervals and /sh-k/’s low-low intervals. Compensation for coarticulation predicts no difference in discriminability between the two pairs because it does not exaggerate differences between the two intervals. [Work supported by NIH.]

Shekgalagari stops and theories of phonological representation

The aim of this paper is to present some aspects of stop consonants, e.g. voice onset time (VOT) and F0 perturbation, and also to present phonological theories of stop representation in order to examine matters of Shekgalagari stop production and phonological representation. Of the phonological representation theories discussed, VOT, the Element theory and Keating's theories of two and three levels of phonological representation as well as the principle of polarization adequately capture phonological contrasts in Shekgalagari. The SPE and the Halle and Stevens (H&S) models are plagued with the problem of generation of more and more features to, if possible, code adequately, the contrasts that exist in Shekgalaga Lwati: A Journal of Contemporary Research Vol. 2 2005: 24-42

The role of acoustic familiarity in the perception of L2 contrasts

Difficulties in L2 perception are often attributed to assimilation of the L2 contrast by an L1 category or to interference by L1‐relevant acoustic cues [Iverson et al. ‘‘Phonetic training with acoustic cue manipulations: A comparison of methods for teaching English /r/‐/p/ to Japanese adults,’’ J. Acoust. Soc. Am. 118, 3267–3278 (2005)]. These approaches assume that cue weights change slowly over developmental time and are relatively constant for short time scales. An alternative view is that cue extraction and weighting are dependent on the overall familiarity of the current sound, with more familiar sounds facilitating more robust cue extraction and better discrimination [e.g., McFadden and Callaway, JEP:HPP, 25, 543 (1999)]. This hypothesis was tested with the perception of English sounds by Japanese adults. The Japanese listeners are insensitive to a cue (F3 onset) when it is part of (putatively unfamiliar) English /ra/ and /la/ sounds. The effect of familiarity was evaluated by fixing the range of F3 onsets and manipulating the rest of each sound in two ways: (1) by changing the transition durations to be more suitable for a moraic rhythm and (2) by changing the F1, F2, and amplitude trajectory to be more like the Japanese /r/. Discrimination improved markedly with the latter manipulation, suggesting that cue sensitivity is conditioned upon overall acoustic familiarity. [Work supported by NIH MH64445.]

Sequential contrast or compensation for coarticulation?

English listeners identify a stop ambiguous between /t/ and /k/ more often as ‘‘k’’ after [s] than [sh] (Mann and Repp, J. Acous. Soc. Am. 69, 548–558, 1981). Judgments shift similarly after a fricative ambiguous between [s] and [sh] when its identity is predictable from a transitional probability bias but perhaps not from a lexical bias (Pitt and McQueen, J. Mem. Lang. 39, 347–370, 1998, cf. Samuel and Pitt, J. Mem. Lang. 48, 416–434, 2003). In replicating these experiments, we add a discrimination task to distinguish between the predictions of competing explanations for these findings: listeners respond ‘‘k’’ more often after [s] because they compensate for the fronting of the stop expected from coarticulation with [s] or because a stop with an F3 onset frequency midway between [t]’s high value and [k]’s low value sounds lower after the [s]’s high-frequency energy concentration. The second explanation predicts listeners will discriminate [s-k] and [sh-t] sequences better than [s-t] and [sh-k] sequences because sequential contrast exaggerates the spectral differences between [s-k]’s high-low intervals and [sh-t]’s low-high intervals and distinguishes them more perceptually than [s-t]’s high-high intervals and [sh-k]’s low-low intervals. Compensation for coarticulation predicts no difference in discriminability between the two pairs because it does not exaggerate differences between the two intervals. [Work supported by NIH.]

Effects of aspiration on fundamental frequency in Taiwanese syllables

The perturbation effect on vowel fundamental frequency (F0) by voiceless aspirated and unaspirated prevocalic obstruents is investigated in Taiwanese. It is well known that F0 is significantly higher after voiceless than voiced stops. However, the perturbation effect caused by aspiration has received much less attention. Twenty-eight minimal pairs contrasting in prevocalic aspiration across three different places of articulation from seven tonal categories in Taiwanese were recorded from two male and two female speakers. An acoustic study was conducted on a total of 1120 syllables. Onset, offset, and mean F0 as well as contours of the tones after the aspirated and unaspirated obstruents were compared. Results indicate that the onset and mean F0 are significantly higher when following aspirated obstruents but no difference was found in offset F0. Tonal contour comparison shows that the F0 raising effect triggered by aspirated obstruents disappears at 50–60<th>% of the tone. Higher larynx and faster airflow rate are posited as the major factors for the higher F0 after voiceless aspirated obstruents. The present results warrant a reconsideration of the traditional theory of aspiration-induced tonal split according to which a tonal split resulted from the lowering of F0 due to aspiration.

Locus equation encoding of stop place: revisiting the voicing/VOT issue

This study explored the consistency of locus equation slopes to serve as phonetic descriptors of stop place in CV utterances across voiced versus voiceless aspirated stops. Using traditional locus equation measurement sites for F2 onsets, voiceless labial and coronal stops had lower locus equation slopes relative to their voiced counterparts, whereas velars failed to show voicing differences. When locus equations were derived using F2 onsets for voiced stops that were measured closer to the stop release burst, comparable to the protocol for measuring voiceless aspirated stops, the slope differences across voicing were greatly minimized. This methodological factor, rather than an underlying phonetic-based explanation, provides a reasonable account for the observed flatter locus equation slopes of voiceless labial and coronal stops relative to voiced cognates.

Perception of place-of-articulation information in natural speech by monkeys versus humans

Four monkeys and 6 humans representing five different native languages were compared in the ability to categorize natural CV tokens of /b/ versus /d/ produced by 4 talkers of American English (2 male, 2 female) in four vowel contexts (/i, e, a, u/). A two-choice "left/right" procedure was used in which both percentage correct and response time data were compared between species. Both measures indicated striking context effects for monkeys, in that they performed better for the back vowels /a/ and /u/ than for the front vowels /i/ and /e/. Humans showed no context effects for the percentage correct measure, but their response times showed an enhancement for the /i/ vowel, in contrast with monkeys. Results suggest that monkey perception of place of articulation is more dependent than human perception on the direction of the F2 onset transitions of syllables, since back-vowel F2s differentiate /b/ and /d/ more distinctively. Although monkeys do not provide an accurate model of the adult human in place perception, they may be able to model the preverbal human infant before it learns a more speech-specific strategy of place information extraction.

Perception of Thai tones in citation form and connected speech

F0 contour shapes of the five tones of Thai differ from citation form to connected speech [Potisuk et al., Phonetica 54, 2242 (1997)]. ‘‘Falling’’ tones have a rise-fall contour in citation, but may be realized as a rise in fluent speech. This study investigates whether Thai listeners can reliably distinguish tones in connected speech, and whether any cues to tone identity remain stable across contexts. In four experiments, ten Thai listeners identified naturally produced and digitally altered tones in a forced-choice task. In experiment one, listeners identified naturally produced ‘‘falling’’ tones with 100% accuracy in citation forms and 96% accuracy in sentences, despite differing contours. In experiments 2–4 (replicating and extending Abramson [Lg&Sp 21, 319325 (1978)]), F0 onset and offset, peak height, and peak alignment were systematically modified on syllables in citation and sentence contexts. In all contexts, tones identified as high or low had an F0 peak or valley aligned to the right edge of the syllable. Mid tones had no F0 inflection. Tones identified as falling or rising necessarily had a pitch inflection at the syllable midpoint (end of the first mora). These findings support the view that tones are aligned with moras in Thai.

The stability of locus equation slopes across stop consonant voicing/aspiration

The consistency of locus equation slopes as phonetic descriptors of stop place in CV sequences across voiced and voiceless aspirated stops was explored in the speech of five male speakers of American English and two male speakers of Persian. Using traditional locus equation measurement sites for F2 onsets, voiceless labial and coronal stops had significantly lower locus equation slopes relative to their voiced counterparts, whereas velars failed to show voicing differences. When locus equations were derived using F2 onsets for voiced stops that were measured closer to the stop release burst, comparable to the protocol for measuring voiceless aspirated stops, no significant effects of voicing/aspiration on locus equation slopes were observed. This methodological factor, rather than an underlying phonetic-based explanation, provides a reasonable account for the observed flatter locus equation slopes of voiceless labial and coronal stops relative to voiced cognates reported in previous studies [Molis et al., J. Acoust. Soc. Am. 95, 2925 (1994); O. Engstrand and B. Lindblom, PHONUM 4, 101–104]. [Work supported by NIH.]

Perception of correlations between acoustic cues in category tuning and speaker adaptation

In English and many other languages, fundamental frequency (f0) varies with voicing such that voiced consonants are produced with lower f0’s than their voiceless counterparts. This regularity robustly influences perception, such that sounds synthesized or spoken with a low f0 are more often perceived as voiced than are sounds with a higher f0. This series of studies exploited these observations to investigate category tuning as a function of incidental exposure to correlations among speech cues and adaptation to speaker idiosyncrasies or accent. Manipulation of f0 across sets of natural speech utterances produced stimulus sets varying in their inherent f0/voicing relationship. Listeners were exposed to these different f0/voicing patterns via spoken word and nonword items in a lexical decision task, and their resulting categorization of ambiguous consonants varying in f0 and voice onset time (VOT) was measured. The results suggest listeners adapt quickly to speaker-specific cues but also remain influenced by more global, naturally occurring covariance patterns of f0 and voicing in English. This pattern contrasts somewhat with studies where idiosyncrasy is represented instead by manipulation of primary, first-order cues to speech sounds, in which listeners are seen to adapt more straightforwardly to the cues they are presented.

The effect of emphatic stress on CV coarticulation

The effect of emphatic stress on CV coarticulation was investigated in the speech of one male and one female native speaker of American English using locus equation slope as a measure of CV coarticulation. Stressed real word C1V2C2 tokens where C1=/b,d,g/ and V2=/i, I, e, ε, æ, u, o, ɔ, a/ were put in carrier sentences with <ub;-.2>the, <ub;-.2>thirty<ubx>, or <ub;-.2>two<ubx> preceding the test word. Each sentence was read three times in a normal manner and three times with emphasis on the test token. This resulted in a total of 486 tokens per speaker (3 stop consonants * 3 V1 contexts * 9 V2 contexts *2 emphasis patterns *3 repetitions). Locus equation slopes were derived by plotting F2 onset of C1 against V2 F2 mid-vowel frequency and fitting a regression line to data points. Consonant closure duration, V2 duration, F0, and amplitude were also measured. Despite a significant increase in the acoustic correlates of emphasis, locus equation slopes remained constant as a function of emphasis and varied as a function of place of articulation. This study provides further evidence of the stability of locus equation slopes as phonetic descriptors of stop place of articulation. [Work supported by NIH.]

Effects of consonant aspiration on Mandarin tones

The influence of consonant aspiration on the F0 contours of tones in Mandarin Chinese was tested in continuous speech by reference to a minimal pair of syllables, /ta/ and /tha/, which differ only in terms of aspiration. It was found that, consonant aspiration affects the fundamental frequency (F0) of the following vowel. The onset F0 of a tone is higher following unaspirated consonants than following aspirated consonants, and the magnitude of the differences are related to the tone itself as well as the preceding tone. The underlying mechanisms of this effect, as well as its interaction with other effects on F0 contours, are discussed.

Effects of speaking rate on second formant trajectories of selected vocalic nuclei.

The effect of speaking rate variations on second formant (F2) trajectories was investigated for a continuum of rates. F2 trajectories for the schwa preceding a voiced bilabial stop, and one of three target vocalic nuclei following the stop, were generated for utterances of the form "Put a bV here, where V was /i/,/ae/ or /oI/. Discrete spectral measures at the vowel-consonant and consonant-vowel interfaces, as well as vowel target values, were examined as potential parameters of rate variation; several different whole-trajectory analyses were also explored. Results suggested that a discrete measure at the vowel consonant (schwa-consonant) interface, the F2off value, was in many cases a good index of rate variation, provided the rates were not unusually slow (vowel durations less than 200 ms). The relationship of the spectral measure at the consonant-vowel interface, F2 onset, as well as that of the "target" for this vowel, was less clearly related to rate variation. Whole-trajectory analyses indicated that the rate effect cannot be captured by linear compressions and expansions of some prototype trajectory. Moreover, the effect of rate manipulation on formant trajectories interacts with speaker and vocalic nucleus type, making it difficult to specify general rules for these effects. However, there is evidence that a small number of speaker strategies may emerge from a careful qualitative and quantitative analysis of whole formant trajectories. Results are discussed in terms of models of speech production and a group of speech disorders that is usually associated with anomalies of speaking rate, and hence of formant frequency trajectories.

Children’s perception of static noise and static formant cues to stop‐consonant place of articulation

Children’s processing strategies appear to favor dynamic cues such as formant transitions as compared to static cues such as F2 onsets and noise bursts. The purpose of this research was to examine children’s perception of place of articulation based only on static cues. Ten children at each of five age levels (3, 4, 5, 6, and 7) and a control group of 10 adults identified synthesized stop consonants [d g] in two vowel contexts [i ɑ]. The synthesis parameters included variations in F2 onsets and stop‐consonant noise bursts. The F2 onsets were either ‘‘appropriate’’ or ‘‘neutral’’ for place of articulation. The noise bursts were either short (10 ms) or long (25 ms). Preliminary data show that the F2 onset is not as salient in children’s perception as in adults’ perception. In addition, children more often than adults categorized neutral F2 onset stimuli as ambiguous indicating stronger category formation in the latter than former groups. The role of noise bursts was more salient in adult perception than chi...

Perception of place‐of‐articulation information by monkeys versus humans

Four monkeys and six humans representing five different native languages were compared in the ability to categorize natural CV tokens of /b/ vs /d/ produced by four talkers of American‐English (two male; two female) in four vowel contexts /i,e,a,u/. A two‐choice left/right procedure was used in which percent correct and response time data were compared between species. Both measures indicated striking vowel context effects for monkeys, but none for humans. Specifically, monkeys performed better for back vowels /a,u/ than front vowels /i,e/. Since back vowels have more distinctive F2 onset transitions differentiating /b/ vs /d/, these results imply that monkey perception is more dependent than human perception on the actual acoustic structure of the syllables. We conclude that humans do not use general mechanisms in place perception, rather they use some sort of special mechanism to eliminate vowel context effects. While monkeys do not provide accurate models of adult humans, they may be able to provide a model of the preverbal human infant before it learns a more speech‐specific adult strategy of place information extraction. [Work supported by NIH.]

An acoustic study of strident fricatives in Mandarin Chinese

Mandarin has a distinction between flat and palatalized post-alveolar fricatives in addition to the alveolar–palatoalveolar contrast, which has been shown to have well-defined acoustic correlates. The goal of the study is to identify the acoustic cues to the place distinction in these strident fricatives and to evaluate how the acoustic properties map onto relevant phonetic features. Acoustic analyses were conducted of 216 fricative-vowel syllables produced by six speakers in three vowel and four tone contexts. Measures were taken of the properties of the frication noise and comparisons between fricative and vowel. The lower frequency limit of the frication noise, spectral peak location, spectral moments, spectral tilt, normalized amplitude, relative amplitude in specific frequency regions, and F2 onset frequency and movement in the adjacent vowel were examined. The results indicate the lower frequency limit of the frication noise is associated with distinct formant regions: F4 or F5 for the alveolar, F3 for the palatalized post-alveolar, and F4 for the flat post-alveolar fricative. Measures that successfully classified all three fricatives include spectral peak location, spectral mean, variance and skewness, amplitude difference between fricative and vowel in the F3 and F4 region, and F2 frequency at vowel onset. [Work supported by NIH.]

Coarticulation and formant transition rate in young children who stutter.

The purpose of this study was to assess anticipatory coarticulation and second formant (F2) transition rate (FTR) of speech production in young children who stutter (CWS) and who do not stutter (CWNS). Fourteen CWS and 14 age- and gender-matched CWNS in three age groups (3-, 4-, and 5-year-olds) participated in a picture-naming task that elicited single-word utterances. The initial consonant-vowel (CV) syllables of these utterances, comprising either bilabial [b m] or alveolar [d n s z] consonants and a number of vowels [i I e epsilon ae u o c aI av],were used for acoustic analysis. To assess coarticulation and speech movement velocity, the F2 onset frequency and F2 vowel target frequency (for coarticulation) and FTR (for speech movement velocity) were computed for each CV syllable and for each participant. Based on these measures, locus equation statistics of slope, y-intercept, and standard error of estimate as well as the FTR were analyzed. Findings revealed a significant main effect for place of articulation and a significantly larger difference in FTR between the two places of articulation for CWNS than for CWS. Findings suggest that the organization of the FTR production for place of articulation may not be as contrastive or refined in CWS as in CWNS, a subtle difficulty in the speed of speech-language production, which may contribute to the disruption of their speech fluency.