F2 Transition Research Articles

Stimulus intensity and fundamental frequency effects on duplex perception.

Duplex perception occurs when part of the acoustic signal is used for both a speech and a nonspeech percept. This phenomenon has been interpreted as evidence of a distinct system for speech perception that precedes other specialized systems of general auditory processing (such as auditory grouping, and perception of pitch, loudness, and timbre). This interpretation was investigated by using an intensity-dependent form of duplex perception with the acoustic pair /da/ and /ga/. The "base" portion of the stimulus, common to both, consisted of the first and second formants and the steady-state portion of the third formant (F3). The F3 transition (either a sinusoid or a true formant), which cued the difference between /da/ and /ga/, was varied in intensity and fundamental frequency (F0). For every subject, the level at which each type of F3 transition was barely audible in the context of the base, i.e., duplex perception threshold, was first established. Next, identification functions were obtained by varying the intensity of the F3 transition relative to each subject's duplex perception threshold. Results revealed that duplex perception thresholds decreased as the F0 of the F3 transition increasingly differed from the base. Also, identification functions showed that, as has been previously demonstrated, the F3 transition contributed to the speech percept over a wide range of intensities and fundamental frequencies. However, as F3 transition intensity increased well above duplex perception threshold, /ga/ identification decreased. Also, both /da/ and /ga/ identification progressively decreased as the F0 of the F3 transition increasingly differed from the base. Contrary to previous duplex perception reports, such findings indicate that both intensity and F0 information is available to the specialized speech perception system. Thus, the computations of the speech perception system and its relation to the general auditory processing systems need to be reexamined.

Reading disability: A deficit in the processing of rapid spectral changes or in phonological coding?

Reading‐impaired children have been said to suffer from an auditory deficit, indexed by difficulties in processing formant transitions [P. Tallal, Brain & Lang. 9, 182–198 (1980)]. Having established that poor readers (n=20) made significantly more errors than good readers (n=20) on discrimination between stop‐vowel syllables, contrasting in initial F2 and F3 transitions (/ba/ and /da/), but not between nonspeech sinewave analogs of F2 and F3, the present study compared the same groups on identification of (i) a /sei‐stei/ synthetic continuum, where F1 transition cued the contrast, and (ii) two hybrid /su–sh/ continua, each icative series by the formant transitions of naturally spoken vowels. Poor readers did not differ from good readers in phoneme boundaries, but did exhibit shallower identification functions, significantly so on the fricative continua. The results do not support the hypothesized auditory deficit: Less consistent discrimination or identification by poor readers on speech tasks but not on the nonspeech control task, suggests phonological rather than auditory difficulties.

Prediction of obstruent spectra from the slope of preceding and following second formant transitions

There is evidence that the slope of second formant (F2) transitions provides an index of the severity of speech mechanism involvement in motor speech disorders [Weismer etal., 1085–1098 (1992)]. From an articulatory perspective, the slope of the F2 transition is an ordinal index of the rate of change in vocal tract configuration. It was hypothesized that, for single words, these F2 slopes would predict quantitative spectral characteristics of adjacent obstruents, under the assumption that the rate of change in vocal tract configuration during a vocalic segment would influence the articulation of a preceding or following obstruent. Quantification of obstruent spectra was achieved by means of spectral moments, which were regressed on F2 transition slopes for a group of normally articulating men and a group with amyotrophic lateral sclerosis, a degenerative motoneuron disease typically associated with dysarthria. Results partially confirmed the hypothesis for both groups of subjects; the findings will be discussed in terms of separate treatment of vocalic and obstruent segments in a model of speech intelligibility. [Work supported by NIH.]

Toward understanding the role of formant transitions for distinctions of stops from glides

Our understanding is incomplete of the properties of vowel formant transitions that contribute to distinctions of voiced stop and glide consonants in speech. Research appears to have established some of the important transition cues for discernment of bilabial synthetic stops versus glides. However, the stop/glide transitions studied have typically been more stylized than those found in natural speech. This investigation examined the importance of transitions to listeners’ identification of initial stops and glides in spoken /CVk/ syllables. Performance was assessed for the stops and glides with progressive deletion of segments from the syllables’ onsets. Bilabial and velar stops and glides as well as alveolar stops were tested in /Ci/, /Copen ayek/, /Cash, ligaturek/ contexts to examine differences in transition use among phoneme environments. Twelve normal‐hearing young adults participated as listeners. In general, when the initial stop bursts were deleted, the F2 transition frequency extent was signifi...

Stimulus intensity and fundamental frequency effects on duplex perception

The simultaneous perception of speech and nonspeech occurs when the intensity of the F3 transition of a three formant synthetic syllable is increased relative to the rest of the syllable (the base). This phenomenon has been interpreted as evidence of a distinct system for speech perception that precedes other specialized systems of general auditory processing [Liberman and Mattingly, Science 243, 489–494 (1990)]. Using F3 transitions with fundamental frequencies different from the base, and referencing their presentation levels to the level at which each F3 transition was barely audible in the context of the base, identification functions were obtained across a wide range of F3 transition intensities. As previously demonstrated, results showed that the F3 transitions contributed to the speech percept over a wide range of intensities and fundamental frequencies. However, at very intense F3 transition levels, /ga/ identification decreased. Also, both /da/ and /ga/ identification progressively decreased as the fundamental frequency of the F3 transition increasingly differed from the base and interacted with intensity. These findings indicate that information from general auditory processing systems is available to the specialized speech perception system. The speech perception system tolerates a range of information from general auditory processing systems before it rejects such information as irrelevant to speech.

Evidence of gestural overlap across speaking rate: Acoustic data

The form of phonetic gestures has been assumed to remain stable across variations in speaking rate. Segmentalization or amount of gestural overlap has been shown to covary with speaking rate. However, the precise relationship between speaking rate and acoustic measures of segmentalization has not been delineated. In the present investigation, speakers produced multiple repetitions of the words ‘‘build,’’ ‘‘dill,’’ and ‘‘gill’’ in carrier phrases at graded, self-selected speaking rates ranging from slow to fast. F2 onset frequency of the vocalic nucleus served as an index of segmentalization and was used to predict the time delay to the onset of the major F2 transition associated with /I1/. Previous work suggests F2 onset frequency is sensitive to degree of gestural overlap in vocalic nuclei preceded by alveolar consonants. Preliminary results suggest (1) longer vocalic nuclei tend to be associated with F2 onset frequencies closer to the preceding consonantal articulatory locus, and (2) the linear relationship between F2 onset and time to target deteriorates above certain speaking durations, suggesting that the basic form of the gesture is altered at slower rates. [Work supported by DC 00319.]

Locus equations during compensatory articulation: A test of the ‘‘orderly output constraint.’’

Previous research [Sussman etal., J. Acoust. Soc. Am. 90, 1309–1325 (1991); 94, 1256–1268 (1993)] showed that frequency of F2 onsets (for syllable-initial voiced stops /bdg/) is a linear function of F2 as measured in the vowel nucleus and the particular linear function relating these parameters is itself a function of stop place of articulation. This linear relationship has been demonstrated cross-linguistically as well as across consonantal manner classes (fricatives, nasals, approximates). The purpose of this study was to determine how altered articulatory kinematics (via bite blocks) affects the linearity and place coding of locus equations. Six speakers produced CV/t/ tokens with initial /bdg/ followed by ten medial vowels. F2 onset (Hz) and F2 midvowel (Hz) were measured via spectrographic, LPC, and FFT spectra. Despite bite blocks locus equations were remarkably linear, and slopes/y intercepts consistently contrasted stop place. A functional role for the acoustic linearity of CV transitions in the output signal, despite gross alterations in articulatory gestures, is offered based on auditory perceptual requirements necessitated by the neural encoding of F2 transitions. These ideas are described within a new theoretical framework—the ‘‘orderly output constraint.’’

Effects of first formant onset frequency on [-voice] judgments result from auditory processes not specific to humans

When F1-onset frequency is lower, longer F1 cut-back (VOT) is required for human listeners to perceive synthesized stop consonants as voiceless. K. R. Kluender [J. Acoust. Soc. Am. 90, 83-96 (1991)] found comparable effects of F1-onset frequency on the "labeling" of stop consonants by Japanese quail (coturnix coturnix japonica) trained to distinguish stop consonants varying in F1 cut-back. In that study, CVs were synthesized with natural-like rising F1 transitions, and endpoint training stimuli differed in the onset frequency of F1 because a longer cut-back resulted in a higher F1 onset. In order to assess whether earlier results were due to auditory predispositions or due to animals having learned the natural covariance between F1 cut-back and F1-onset frequency, the present experiment was conducted with synthetic continua having either a relatively low (375 Hz) or high (750 Hz) constant-frequency F1. Six birds were trained to respond differentially to endpoint stimuli from three series of synthesized /CV/s varying in duration of F1 cut-back. Second and third formant transitions were appropriate for labial, alveolar, or velar stops. Despite the fact that there was no opportunity for animal subjects to use experienced covariation of F1-onset frequency and F1 cut-back, quail typically exhibited shorter labeling boundaries (more voiceless stops) for intermediate stimuli of the continua when F1 frequency was higher. Responses by human subjects listening to the same stimuli were also collected. Results lend support to the earlier conclusion that part or all of the effect of F1 onset frequency on perception of voicing may be adequately explained by general auditory processes.(ABSTRACT TRUNCATED AT 250 WORDS)

F2 Transitions During Sound/Syllable Repetitions of Children Who Stutter and Predictions of Stuttering Chronicity

The purpose of the present study was to examine the relationships between second formant (F2) transitions during the sound/syllable repetitions (SSRs) of young children who stutter and their predicted chronicity of stuttering. Subjects were 13 youngsters who stutter, who were divided into two groups based on their predicted chronicity of stuttering as measured by the Stuttering Prediction Instrument (SPI; Riley, 1984): a high-risk group, consisting of 7 boys (mean age = 50.6 months), and a low-risk group, consisting of 5 boys and 1 girl (mean age = 48.5 months). Each child was audio/videotape-recorded during a 30-minute conversational interaction with his or her mother. Ten SSRs per child were acoustically analyzed to identify differences in F2 transitions between the repeated (stuttered) and fluent (nonstuttered) portions of the words. Present findings are consistent with those of Stromsta (1965, 1986), who reported that children who stutter produce F2 transitions during stuttering that are nonmeasurable or missing or that differ in direction of movement from fluent transitions. However, there were no significant between-group differences in the frequency of occurrence of these "abnormal" F2 transitions, findings that are apparently inconsistent with Stromsta's results. The remaining measurable F2 transitions showed no significant between-group differences in the mean differences between stuttered and fluent F2 transitions for onset and offset frequencies, transition extents, and transition rates. Within both groups, significant positive correlations were found between stuttered and fluent F2 transitions for all acoustic measures except for transition durations, which were not significantly correlated for either high-risk or low-risk subjects. Within the low-risk group, stuttered F2 transitions were typically shorter than fluent transitions. Findings were taken to suggest that some elements of sound or segment prolongation may be present within the SSRs of children who stutter and who are considered to be at high risk for continuing to stutter, indicating that further study of selected aspects of F2 transitions during stuttering may provide useful clinical information for predicting the likelihood that a child will continue to stutter.

The perceptual weighting of acoustic cues changes with linguistic experience

Previous studies found that children’s judgments of syllable-initial /s/ and /∫/ are more related to the vocalic F2 transition and less related to the fricative-noise spectrum than are adults’ judgments [e.g., Nittrouer, J. Phon. 20, 351–382 (1992)]. Such results have led to a model of speech development proposing that children’s weighting of acoustic cues changes as they gain linguistic experience. The present study tested two requisites of that model, namely that the perceptual weighting of acoustic cues must be flexible and cannot simply reflect a listener’s auditory sensitivities. Adults and 3-year olds participated in two tasks: Identification tasks, using synthetic fricative noises and either natural or synthetic vocalic portions; and discrimination tasks, measuring sensitivity to fricative-noise spectrum and F2 transition. Identification tasks showed the same age-related differences found in earlier studies when natural vocalic portions were used, but these age effects were reduced when synthetic vocalic portions were used. Discrimination tasks showed slightly larger difference thresholds for both the fricative noise and the F2 transition for children than for adults, but this age effect could not explain the age effect on the weighting of those cues for identification. It was concluded that the weighting of acoustic cues is flexible both for adults and for children, and that age-related differences in the weighting of the cues to /s/ and /∫/ are not explained by age-related differences in auditory sensitivities. [Work supported by NIH.]

Relationship between identification errors and spectral details of naturally produced diphthongs

In an earlier study [A. K. Nabelek et al., J. Acoust. Soc. Am. 92, 1228–1246 (1992)], identification errors were collected for 15 English vowels, monophthongs, and diphthongs. The vowels were produced in a /b–t/ context by six male talkers who uttered five tokens of each vowel. There were 20 normal‐hearing and 20 hearing‐impaired subjects. Listening conditions were quiet, noise, and reverberation. Large differences in number of errors were observed among the 30 tokens of each vowels. The goal of the current study was to determine relationships between the spectral details of the diphthongs /open ayei, open ohi, open ayeu/ and identification errors. Spectral analyses indicated that all three diphthongs had two distinct segments. In the first segment F1 and F2 were relatively steady state and in the second segment F2 changed over time. For each token of each diphthong, frequencies and levels of F1 and F2 were traced along the duration of the diphthong. The results of this analysis seem to indicate a relationship between the number of identification errors for these diphthongs and the relative levels of the F2 transition segment. Examples of highly identifiable and high confusable diphthongs will be shown. [This work was supported by NIDCD.]

Attentional effects on phonetic encoding of acoustic cues

The role of attention in the phonetic encoding of acoustic cues was studied by examining the effect of a nonspeech distractor task on listeners’ classification of the speech sounds /ra/ and /la/. Consistent with previous results [Gordon etal., Cog. Psychol. 25 (1993)] it was found that the impact of a relatively strong cue (F3 transition) on phonetic classification was reduced when listeners could not devote full attention to the speech sound. In contrast with previous results, the impact of a relatively weak cue (an aperiodic burst during the formant transition) was also reduced when full attention was not possible. This result conflicts with models in which the importance of attention in phonetic encoding is greater for acoustic cues whose phonetic importance is very great. It suggests that the role of attention in phonetic encoding depends on specific characteristics of acoustic cues.

Effects of discontinuous F3 transitions on phonetic identification and fusion

According to general perceptual principles, figural goodness and feature integration of stimuli should reflect the continuity between components; in CV syllables this continuity might be between phoneme-distinguishing glide (formant transition) and vowel formant. Two experiments examined continuity effects on identification and fusion for sinewave syllables based on common CV parameters. Experiment 1 provided a perceptual basis for equating isolated glides and stationary tones. Independent of glide direction, perceptual matches tended to be based upon temporally adjacent portions of the stimuli. Using matched stimuli, experiment 2 evaluated the relative contribution to phonetic identification of (1) frequency continuity of transition and formant and (2) transition direction. Subjects discriminated the original syllables from binaural and dichotic syllables containing either continuous or discontinuous F3 transitions. Identification was influenced more by transition direction than by goodness of physical continuity. Furthermore, very different patterns of results were obtained for binaural conditions and dichotic conditions in which responses indicated fusion. The nature of information provided by transitions and operation of schema are discussed. [Work supported by NSF and AFOSR.]

Comparison of human and monkey sensitivity to the English liquid /ra–la/ contrast

This study compared human and monkey sensitivity to the English liquid /ra–la/ contrast using a low-uncertainty repeating-standard discrimination procedure. A synthetic continuum was constructed by varying both a spectral cue in the third formant (F3 onset frequency) and a temporal cue in the first formant (steady-state period before the F1 transition into the vowel). In experiment 1, differential sensitivity was measured using both the /ra/ and /la/ ends of the continuum as standards. Humans were less sensitive than monkeys to variation at the /ra/ end, but equally sensitive at the /la/ end. In experiment 2, subjects were presented with pairwise comparisons from the continuum to discriminate. Human discrimination was best at the identified phoneme boundary near the center of the continuum, while monkey discrimination was best inside the /ra/ category. These results indicate qualitative differences between human and monkey discrimination of this continuum: categorical-like for humans, and noncategorical for monkeys. [Work supported by NIDCD.]

A cross-linguistic validation of locus equations: Are there ‘‘phonetic hot spots’’ in CV acoustic space?

A previous study [Sussman etal., J. Acoust. Soc. Am. 90, 1309–1325 (1991)] of American English CV coarticulation showed a remarkably linear relationship between onset frequencies of F2 transitions (plotted along the ordinate) in relation to midvowel target frequencies (plotted along the abscissa) in CVC tokens with initial /bdg/ followed by ten medial vowels. Slope and y intercept of regression functions (‘‘locus equations’’) fitted to these coordinates systematically varied as a function of stop place. Discriminant analyses, using as predictor variables, slope and y intercept, yielded 100% correct classification of stop place categories. Locus equations provide a systematic lawfulness to coarticulatory variation and a relationally invariant phonetic index for stop place classification. The present study extends the metric to three additional languages−Thai, Cairean Arabic, and Urdu. Resulting scatterplots were extremely linear and varied as a function of stop place. Plotting slope Xy intercept yields a derived map of CV phonetic space with which to relationally compare 2, 3, and 4-stop place languages. Within-language stop place contrasts were consistly divergent in CV space. Variability of labial, alveolar/dental, and velar coordinates across five languages showed fairly broad clustering of stop place categories, rather than narrowly focused ‘‘phonetic hot spots.’’ The data are discussed in relation to quantal and adaptive dispersion theories. [Work supported by NSF.]

Effects of first formant onset frequency on voicing judgments result from auditory processes not specific to humans

When Fl onset frequency is lower, a longer F1 cutback (VOT) is required for human listeners to perceive synthesized stop consonants as voiceless. K. R. Kluender [J. Acoust. Soc. Am. 90, 83–96 (1991)] found comparable effects of F1 onset frequency on the ‘‘labeling’’ of stop consonants by Japanese quail (Coturnixcoturnixjaponica) trained to distinguish stop consonants varying in F1 cutback. In that study, CV’s were synthesized with natural-like rising F1 transitions. Consequently, endpoint training stimuli (with 5- or 65-ms F1 cutback) differed in the onset frequency of F1 because a longer cutback resulted in a higher F1 onset. In order to assess whether earlier results were due to auditory predispositions or due to animals learning the covariance between F1 cutback and F1 onset frequency, the present experiment was conducted with synthetic continua having either a relatively low (375-Hz) or high (750-Hz) constant-frequency F1. Eight birds were trained to respond differentially to endpoint stimuli from three series of synthesized /Ca/’s varying in duration of F1 cutback. Second and third format transitions were appropriate for labial, alveolar, or velar stops. Despite the fact that there was no opportunity to use experienced covariation of F1 onset frequency and F1 cutback, quail exhibited reliably shorter ‘‘labeling’’ boundaries (more voiceless stops) for intermediate stimuli of the continua when F1 frequency was higher. Human performance with the same stimuli was like that for the birds. These results support the conclusion that the effect of F1 onset frequency may be adequately explained by general auditory processes. [Work supported by NIDCD/NIH Grant No. DC-00719.]

Transition duration effects on place perception in hearing-impaired listeners.

The purpose of this study was to determine the contribution of transition duration in signaling an alveolar/velar place-of-articulation contrast to listeners with mild-to-moderate sensorineural hearing loss. In the experiments reported here, F1, F2, and F3 transitions were varied between values appropriate for [deh] and [geh]. These were combined with transition durations ranging from 20 to 50 ms in 5-ms steps. In three separate identification tasks, listeners were presented with: (1) all levels of formant pattern varied orthogonally with all transition duration values; (2) all levels of formant pattern combined with an ambiguous duration value; and (3) all levels of transition duration combined with an ambiguous formant pattern. A trend was observed for hearing-impaired listeners to identify stimuli with shorter transition durations as [deh] in the absence of other unequivocal cues for place of articulation. Overall performance of hearing-impaired listeners was more variable than that of normal listeners, predicted in part by the results of Summerfield etal. [Speech Commun. 4, 213–229 (1985)], in which hearing-impaired listeners showed a decreased ability to identify place of articulation from transition information alone. Results of these experiments will be discussed in terms of current views of temporal processing in hearing-impaired listeners. [Work supported by NIH.]

The phonetic integration of speech and non-speech sounds: effects of perceived location.

The third-formant (F3) transition of a three-formant /da/ or /ga/ syllable was extracted and replaced by sine-wave transitions that followed the F3 centre frequency. The syllable without the F3 transition (base) was always presented at the left ear, and a /da/ (falling) or /ga/ (rising) sine-wave transition could be presented at either the left, the right, or both ears. The listeners perceived the base as a syllable, and the sine-wave transition as a non-speech whistle, which was lateralized near the left ear, the right ear, or the middle of the head, respectively. In Experiment 1, the sine-wave transition strongly influenced the identity of the syllable only when it was lateralized at the same ear as the base (left ear). Phonetic integration between the base and the transitions became weak, but was not completely eliminated, when the latter was perceived near the middle of the head or at the opposite ear as the base (right ear). The second experiment replicated these findings by using duplex stimuli in which the level of the sine-wave transitions was such that the subjects could not reliably tell whether a /da/ or a /ga/ transition was present at the same ear as the base. This condition was introduced in order to control for the possibility that the subjects could have identified the syallables by associating a rising or falling transition presented at the left ear with a /da/ or /ga/ percept. Alternative suggestions about the relation between speech and non-speech perceptual processes are discussed on the basis of these results.

Young infants’ sensitivity to formant transitions.

Infant discrimination of pairs of vowels, CV and VC syllables was studied with Swedish subjects whose mean age was about 100 days. The data were obtained using the high-amplitude sucking paradigm. Eighty-seven subjects were tested in four different sessions, during which the following contrasts were studied: [ɑ ] vs [a], [bɑ ] vs [dɑ ], [ba] vs [bɑ ], and [ɑ b] vs [ɑ d]. The contrasts were obtained by a 2-Bark shift of the F2 frequency. When applied to the vowel’s F2, this shift resulted in two distinct vowels. When applied to the onset (or offset) of F2 transitions, it generates a /b/ vs /d/ contrast. To control for learning effects due to the repeated measures procedure, the overall experiment is counterbalanced. Each subject was randomly assigned to one of the possible 24 permutations of the four experimental conditions. The results indicate a more reliable discrimination for stimuli in which the relevant information was carried by initial or final formant transitions.

Age-related changes in articulation of /b/ and /w/.

The present study has examined the effect of the age of the speaker on the duration of prevoicing, release, and transition in the articulated phonemes /b/ and /w/. Thirty normal subjects, belonging to three age groups: 20–30, 50–60, and 70–80 years old, were asked to pronounce the syllables /wa/ and /ba/. It was found that prevoicing in both phonemes becomes significantly longer with age; however, duration of the release segment is not influenced by age. Results demonstrated that speakers in the three age groups differed significantly in the duration of the articulated /w/ transition (the elderly produced a significantly shorter transition than the younger subjects). Moreover, analyses demonstrated a relationship between the duration and the difference between offset and onset frequencies of F2 transition of the /w/ phoneme. This is probably due to age-related changes in tongue and jaw movements during /w/<ep;j>articulation. Moreover, the perceptual results indicated that the boundary of phonemic contrast /b–w/ was shifted significantly toward shorter transition durations as a function of age.