Reiterant Speech Research Articles

Reiterant speech as a test of non-native speakers’ mastery of the timing of French

The reiterant speech of ten native speakers of French was analyzed to develop baseline measures for syllable and consonant/vowel timing for a series of two-, three-, four-, and five-syllable French words spoken in isolation. Ten native speakers of English, who learned French as a second language, produced reiterant versions of both the French words and a comparable set of English words. The native speakers of English were divided into two groups on the basis of their second language experience. The first group consisted of four university-level teachers, who were relatively experienced learners of French, and the second group of six less experienced learners of French. The French reiterant imitations of the two groups of native speakers of English were compared to the native French speakers' productions. The timing patterns of the experienced group of non-native speakers did not differ significantly from those of the native French speakers, whereas there was a significant difference between these two groups and the group of six less experienced second-language learners. Deviations from the French baseline measures produced by the less experienced group are discussed in terms of the influence of the timing patterns of English and the literature on a sensitive period for second language acquisition.

Fundamental frequency and perceived prominence of accented syllables

Perceived prominence differences between accented words have been accounted for in terms of fundamental frequency (F0) variations. In one theory prominence is related to the relative magnitude of F0 changes, and in another theory prominence is related to relative frequencies of F0 maxima. Two experiments with the same group of subjects contrasted the two theories. The experiments used reiterant speech consisting of "maMAmamamaMAma" utterances with synthesized F0 contours, with F0 maxima on the second and penultimate syllables (P1 and P2, respectively). In one experiment the utterances had no baseline declination, and in the other experiment the slope of the baseline was varied. In both experiments, subjects performed an adjustment task under two different instructions: They were instructed to adjust P2 so that it had the same pitch or prominence as P1. The results show that both variables (instruction and presence or absence of baseline declination) affect the adjustments of P2. It is concluded that neither theory gives a satisfactory account of the results, and that perceived prominence is related in a complex way to the range of F0 values employed.

Fundamental frequency and perceived prominence of accented syllables

In natural speech, accented words may differ as to the degree of perceived prominence. At the acoustic level, two aspects of fundamental frequency (F0) variation may be responsible for the perceived prominence differences: the magnitude of F0 changes and the relative frequencies of F0 maxima. Two experiments, with the same group of subjects, addressed the question of which aspect of F0 better predicts perceived prominence. Both experiments used reiterant speech with synthesized F0 contours. The speech materials consisted of ‘mamamamamamama’ utterances with F0 maxima on the second and penultimate syllables (“P1” and “P2,” respectively). In one experiment, subjects adjusted the frequency of P2 so that it was judged to have the same pitch as P1, for different rates of baseline declination. In the second experiment, subjects adjusted P2 so that it was judged to have the same prominence as P1, again for different declination rates. The results to be presented are relevant for refining the theory of pitch accentuation. For instance, if perceived prominence is predictable from F0 maxima, both experiments should give the same results.

The cricothyroid muscle in voicing control.

Initiation and maintenance of vibrations of the vocal folds require suitable conditions of adduction, longitudinal tension, and transglottal airflow. Thus manipulation of adduction/abduction, stiffening/slackening, or degree of transglottal flow may, in principle, be used to determine the voicing status of a speech segment. This study explores the control of voicing and voicelessness in speech with particular reference to the role of changes in the longitudinal tension of the vocal folds, as indicated by cricothyroid (CT) muscle activity. Electromyographic recordings were made from the CT muscle in two speakers of American English and one speaker of Dutch. The linguistic material consisted of reiterant speech made up of CV syllables where the consonants were voiced and voiceless stops, fricatives, and affricates. Comparison of CT activity associated with the voiced and voiceless consonants indicated a higher level for the voiceless consonants than for their voiced cognates. Measurements of the fundamental frequency (F0) at the beginning of a vowel following the consonant show the common pattern of higher F0 after voiceless consonants. For one subject, there was no difference in cricothyroid activity for voiced and voiceless affricates; in this case, the consonant-induced variations in the F0 of the following vowel were also less robust. Consideration of timing relationships between the EMG curves for voiced and voiceless consonants suggests that the differences most likely reflect control of vocal-fold tension for maintenance or suppression of phonatory vibrations. The same mechanism also seems to contribute to the well-known difference in F0 at the beginning of vowels following voiced and voiceless consonants.

Kinematic analysis of articulatory declination

Previous investigations of declination—the progressive reduction of some measure over the course of an utterance—have focused primarily on F0, with occasional mention of acoustic amplitude. Recently, F0 declination has been shown to track a decline in subglottal pressure. To our knowledge, whether declination occurs in other subsystems associated with speech production has not been investigated. One such subsystem currently being examined for speakers of English, French, and Japanese is that of supralaryngeal articulator motion. Kinematic variables of displacement, duration, and peak velocity, associated with motion of the lower lip/jaw complex during reiterant speech, were tested for declination effects. Initial findings show that there is articulatory declination for some speakers, but there is much interspeaker variability when kinematic variables are examined individually. However, as demonstrated previously for stress and speaking rate contrasts, much of the spatiotemporal variability may be accounted for when the relations among kinematic variables are considered, thereby making it easier to determine whether articulatory declination is present. Furthermore, for the data examined thus far, articulator motion appears to adhere to the constraints of a damped, second-order dynamical system, whether articulatory declination is observed or not. [Work supported by NIH.]

Voice source variations during consonant-vowel transitions

Studies of the voice source during speech have been concerned mostly with variations in fundamental frequency and amplitude. However, the source variations also comprise the spectral characteristics and the harmonics-to-noise ratio. The present study reports detailed measurements of voice source properties during transitions between consonants and vowels. Oral air flow, oral air pressure, and transillumination were recorded from two subjects producing reiterant speech with the vowel /æ/ and different voiced and voiceless consonants; the flow signal was inverse-filtered to obtain an estimate of the glottal pulse. Results indicate a breathy phonation at the release of voiceless consonants as indicated by an open quotient close to 1. Peak flow during each glottal pulse is thus high at voicing onset following a voiceless consonant and decreases, with an undershoot, during the first 20 pitch periods. The source pulse is also skewed to the left during the first pitch periods following voiceless consonants. Source variations following voiced consonants are less pronounced. [Work supported by NINCDS Grants NS-13617 to Haskins Laboratories.]

Acoustic correlates of breathiness: First harmonic amplitude, turbulence noise, and tracheal coupling

A selected sample of reiterant speech has been collected from ten female speakers and six male controls in order to quantify acoustic correlates of perceived breathiness of the female voice, and to contrast these measures with comparable data from males. Two sentences with differing stress patterns were spoken by replacing each syllable by [hV] and by [ʔV], where V = [a,i,ae,o,rr]. Detailed analysis of the [a] data reveals (1) wide variation in the strength of the first harmonic (relative to first formant amplitude), with an average increase of about 6 dB for females relative to males; (2) a greater tendency for the third formant to be excited by noise rather than voicing harmonics in the female population; and (3) indications of tracheal poles and zeros in the spectra of vowels adjacent to voiceless consonants in utterances produced by both genders. These three measures of breathiness tend to be greatest in unstressed syllables and toward the end of an utterance. The acoustic data will be correlated with perceived breathiness of utterances from individual talkers in a preliminary attempt to determine which measure is perceptually most salient; future research using synthesis will systematically investigate perceptual salience. [Work supported, in part, by an NIH grant.]

Using metrical phonology to synthesize speech

In this experiment synthesized reiterant speech is compared to natural reiterant speech. Using multiple linear regression, equations are obtained which take grammatical information (e.g., syllable weight, word boundary, lexical stress) of natural speech syllables as input and produce as output information sufficient to predict the timing of the corresponding reiterant speech in nonphrase final position. A corpus of 35 phrases was spoken reiterantly by three speakers, using the syllable sa as the reiterant syllable. Measurements of the duration of the reiterant segments were made automatically from the digitized speech and constitute the final dependent variable. Each syllable is grammatically described as a string of binary variables determined from metrical phonology [Liberman and Prince (1977)]. After calculating the predicted segmental durations of a reiterant phrase, the original speech was modified to fit these predictions using the waveform editor of ILS. The resulting synthesized speech is then jud...

Interarticulatory timing and single articulator velocity-displacement relationships in English stress pairs

Evidence has been brought forward that there exists a direct timing relationship between the articulator movements associated with vowel production and those associated with consonant production. Tuller and others have interpreted such results as evidence for a model of speech production that posits a metrical relationship between vowels and their coproduced consonantal neighbors [cf. Harris et al., in Invariance and Invariability in Speech Production, edited by J. Perkell and D. Klatt (Erlbaum, Hilldale, NJ, 1986); and Tuller et al., J. Exp. Psychol. 9, 829–833 (1983)]. A more strictly segmental ordering of speech events could be posited to handle such data. The difference between these models can be brought to light by making reference to statistical correlations between the timing latencies of various articulatory events [cf. K. Munhall, J. Acoust. Soc. Am. 78, 1548–1553 (1985)]. This study investigates the effects of stress pattern and syllable structure on articulator timing, using x-ray microbeam traces of jaw, tongue blade, and lip movement. Unlike in earlier studies, real English words were used. Each token was placed in two contexts, natural sentences and short, frame sentences. The effects of these two environments are also to be noted. The traces will also be used to replicate the velocity-displacement relationship in the movement of the jaw, found in earlier studies of reiterant speech [Kelso et al., J. Acoust. Soc. Am. 77, 266–280 (1985)].

A dynamic analysis of reiterant speech production in three languages

Previously, supralaryngeal articulatory motion in reiterant speech produced by speakers of English and Japanese in terms of an underlying dynamical model [Kelso et al., 1. Acoust. Soc. Am. 77, 266–280 (1985); Vatikiotis‐Bateson and Kelso, J. Acoust. Soc. Am. Suppl. 1 78, S38 (1985)] has been analyzed. Here, the analysis is extended to French, a language whose temporal structure is believed to be different from that of English or Japanese, and a preliminary comparison of movements of the lower‐lip/jaw complex for speakers of the three languages is presented. The cross‐language comparison is based on articulatory movement data from 13 speakers (five English, five Japanese, and three French). Analysis of reiterant /ba/ and /ma/ productions reveals that French speakers make language‐specific linguistic (e.g., stress) and performance (e.g., speaking rate) distinctions that are qualitatively similar in terms of kinematic relations to those observed for Japanese and English. Quantitative differences due to the intrinsic temporal properties of each language are also evident (e.g., in the slope of the peak velocity‐displacement relation). It is suggested that the same abstract dynamical model, appropriately parameterized for each language, can characterize the observed articulatory behavior. [Work supported by NINCDS and ONR.]

A dynamic analysis of reiterant speech production in Japanese

Previously, we have modeled supralaryngeal articulatory motion in the production of English reiterant speech in terms of underlying dynamical constraints on movement gestures [Kelso et al., J. Acoust. Soc. Am. 77, 266–280 (1985)]. Here, we extend the model to Japanese, a language whose temporal structure is believed to be based on the “mora” and thus very different from English. Preliminary analysis of acoustic lip and jaw movement, and electroglottographic data from three Japanese talkers reveal production differences between the reiterant renditions of multimora and single mora syllables. Multimora CVC syllables are characterized by a glottally truncated vowel followed by a lengthened silent period prior to voicing onset for the following syllable, e.g., a word like 〈motte〉 is realized reiterantly as [ba?ba]. Thus the timing of the glottal gesture appears specific to the mora structure of Japanese, while upper articulator motion, which does not differ for the two syllable types, can be shown to adhere to dynamic constraints hypothesized for movement gestures in general. [Work supported by NINCDS and ONR.]

A qualitative dynamic analysis of reiterant speech production: phase portraits, kinematics, and dynamic modeling.

The departure point of the present paper is our effort to characterize and understand the spatiotemporal structure of articulatory patterns in speech. To do so, we removed segmental variation as much as possible while retaining the spoken act's stress and prosodic structure. Subjects produced two sentences from the "rainbow passage" using reiterant speech in which normal syllables were replaced by /ba/ or /ma/. This task was performed at two self-selected rates, conversational and fast. Infrared LEDs were placed on the jaw and lips and monitored using a modified SELSPOT optical tracking system. As expected, when pauses marking major syntactic boundaries were removed, a high degree of rhythmicity within rate was observed, characterized by well-defined periodicities and small coefficients of variation. When articulatory gestures were examined geometrically on the phase plane, the trajectories revealed a scaling relation between a gesture's peak velocity and displacement. Further quantitative analysis of articulator movement as a function of stress and speaking rate was indicative of a language-modulated dynamical system with linear stiffness and equilibrium (or rest) position as key control parameters. Preliminary modeling was consonant with this dynamical perspective which, importantly, does not require that time per se be a controlled variable.

Some fundamental frequency perturbations in a sentence context

In the present investigation fundamental frequency perturbations in vowel segments owing to preceding and following stop consonants were studied by means of reiterant speech that is assumed to simulate the conditions in connected speech with a normalized syllable structure. The perturbations of the fundamental frequency contour initially after the voice onset of a vowel were found to be larger at low frequencies than high and in the range 10–15 Hz at low frequencies for a male subject. At the end of the vowel segment followed by a voiceless stop consonant there is a tendency for lowering the fundamental frequency contour. The perturbations observed seem to be possible to explain by physiological factors and not support any reorganization of the motor command in order to minimize the perturbing effects.

A dynamic analysis of reiterant speech production

Speaking, like other biological activities, emerges from a neuromuscular basis of high dimensionality as a coherent pattern of articulator motions. What form, if any, does such coherence take and how might it be rationalized? Conventional analyses of articulatory motion are not always informative. Here we use the tools of qualitative dynamics, particularly phase portrait techniques [cf. J. A. S. Kelso and E. V. Bateson, J. Acoust. Soc. Am. Suppl. 1 73, S67 (1983)] to analyze articulatory gestures from an experiment in which speaking rate and stress are naturally varied. We show that the strong relationship between a gesture's amplitude and peak velocity may be accounted for when movements are analyzed in an action (energy × duration) versus amplitude performance space. The action for a given gesture (associated with consonant‐vowel and vowel‐consonant transitions) and for a given speaker scales to (amplitude)2. This finding suggests (1) that stiffness is a key dynamic parameter of the articulatory system, and (2) that an elastic similarity principle may underlie the articulatory behavior of different speakers. [Work supported by NIH, BRSG, and ONR.]

Syllable Duration in Stress-, Syllable- and Mora-Timed Languages

Abstract Building on the results of a previous study, reiterant speech is used to test a number of durational characteristics of a stress-timed language (English), a syllable-timed language (Spanish) and a mora-timed language (Japanese). The results indicate that moras and stress feet do influence syllable duration in their respective languages, Japanese and English. Spanish, however, seems to show no language-specific durational characteristics which could be claimed to be indicative of syllable timing.

On the cyclical basis of speech production

Though certain languages are hypothesized to possess an inherently rhythmical structure, the basis of any temporal regularity in speech production remains elusive. Efforts to identify rhythmical properties in the speech signal have been plagued by the influence of segmental variation on the measurement process. Here we remove such variation as much as possible but do not annihilate the spoken act's stress and prosodic structure. Subjects produced sentences from the rainbow passage in a reiterant speech mode in which normal syllables were replaced by /ba/ or /ma/. Infrared LEDs were placed on the jaw and lips and monitored using a modified SELSPOT optical tracking system. When pauses marking syntactic boundaries were removed, a high degree of rhythmicity was observed revealed by a periodicity in peak-to-peak jaw displacement of 5–6 Hz and small coefficients of variation (SD of jaw period/mean period). Though the cycling was not perfectly isochronous, it nevertheless exhibited orbital stability on the phase portrait (jaw displacement versus velocity). Such stability regimes are characteristic of nonlinear, limit cycle clocking modes which seem to underlie a variety of motor activities, and may constitute the fundamental “grid” within which segmental events must cohere. [Work supported by NINCDS, BRSG, and ONR.]

Reiterant speech: an acoustic and perceptual validation.

Reiterant speech, or nonsense syllable mimicry, has been proposed as a way to study prosody, particularly syllable and word durations, unconfounded by segmental influences. Researchers have shown that segmental influences on durations can be neutralized in reiterant speech. If it is to be a useful tool in the study of prosody, it must also be shown that reiterant speech preserves the suprasegmental duration and intonation differences relevant to perception. In the present study, syllable durations for nonreiterant and reiterant ambiguous sentences were measured to seek evidence of the duration differences which can enable listeners to resolve surface structure ambiguities in nonreiterant speech. These duration patterns were found in both nonreiterant and reiterant speech. A perceptual study tested listeners' perception of these ambiguous sentences as spoken by four "good" speakers--speakers who neutralized intrinsic duration differences and whose sentences were independently rated by skilled listeners as good imitations of normal speech. The listeners were able to choose the correct interpretation when the ambiguous sentences were in reiterant form as well as they did when the sentences were spoken normally. These results support the notion that reiterant speech is like nonreiterant speech in aspects which are important in the study of prosody.

Another look at reiterant speech in prosody research

Reiterant speech or nonsense syllable mimicry [Liberman and Streeter, J. Acoust. Soc. Am. 63, 231–233 (1978)] is useful for studying prosodic phenomena if it can be shown to preserve the suprasegmental duration and intonation differences relevant to perception, while neutralizing the intrinsic segmental duration differences normally found in speech. Syllable durations for normal and reiterant ambiguous sentences were measured to seek evidence of the duration differences which can enable listeners to resolve surface structure ambiguities in normal speech. These duration patterns were found in both normal and reiterant sentences produced by “good” speakers—speakers who neutralized intrinsic duration differences and whose sentences were independently rated by skilled listeners as good imitations of normal speech. Further, listeners were able to choose the correct interpretation when the ambiguous sentences were in reiterant form as well as they did when the sentences were spoken normally. [Supported by NIH and the University of Minnesota.]

Prosodic aspects of American English speech rhythm

We studied American English speech rhythm using reiterant speech with the prosody of normal speech but without its segmental variations. Syllable duration was strongly influenced by stress and by word and phrase boundaries, but negligibly by other nonfinal positions within words and phrases. Word duration increased linearly with word size, and varied with stress pattern for words of constant size. Metrical foot duration increased approximately linearly with foot size. Feet of the same type and size were of constant duration.

Prosodic Aspects of American English Speech Rhythm

Abstract We studied American English speech rhythm using reiterant speech with the prosody of normal speech but without its segmental variations. Syllable duration was strongly influenced by stress and by final positions in words and phrases, but negligibly by nonfinal positions within words and phrases. Word duration increased linearly with word size, and varied with stress pattern for words of constant size. Metrical foot duration increased approximately linearly with foot size. Feet of the same type and size were of constant duration.