Consonant Release Research Articles

Selective adaptation for acoustic cues of voicing in initial stops

Recent experiments have indicated that perceptual adaptation effects can be obtained for a series of CV syllables (e.g. [ba–[pha]) varying along the acoustic dimension of Voice Onset Time (VOT), defined as the interval between the onset of the consonant release burst and the onset of voicing. These adaptation effects have led researchers to postulate the existence of a small set of linguistic feature detectors, each sensitive to a restricted range of Voice Onset Time values. However, another acoustic cue for voice-voiceless distinctions among the initial stop consonants, namely the presence or absence of significant formant transitions following the onset of voicing, covaried with absolute VOT in each of the previous adaptation experiments. In this study, differential adaptation effects are obtained along a VOT test dimension by presenting adapting stimuli which contain different amounts of formant transitions after voicing onset but which contain the same absolute VOT. The results indicate that the original adaptation effects are partly attributable to the presence or absence of the formant transitions after voicing onset, rather than to VOT per se. The present results may be attributable to adaptation of a detector sensitive to a weighted combination of the two hypothetical cues of VOT and the duration of voiced transitions.

Interjected sounds in deaf children's speech

Recordings and phonetic transcriptions of deaf children's speech were examined for nonstandard interjected elements associated with consonant production. Vowel release of both stop consonants and fricatives is found in word-final position and between abutting or normally blended sounds, particularly where assimilation ordinarily occurs, such as between stops or between a stop and fricative. A lateral /1/ is also used as a release. Stop positions are arrived at too slowly, introducing a fricative of similar place and voice. A stop released with overeffort produces an affricate. A fricative may be preceded by a stop of similar place and voice. Cessation of nasality without movement from a position produces a plosive following the nasal sound. Extra nasals are inserted before or after a stop in the same place. The substitutions in deaf children's speech often result from undershoot of the articulatory target. Some interjected elements apparently represent overshoot, which may be due to conscious effort at articulatory placement, while others stem from slowness of movement of the articulators.

Voice-timing perception in Spanish word-initial stops

We have shown in earlier work that speakers of Spanish differentiate voiced /bdg/ from voiceless /ptk/ in word-initial position by means of voice onset time (VOT), the temporal relation between the onset of glottal pulsing and acoustic features of supraglottal articulation. More recently, we demonstrated the perceptual efficacy of VOT for the three categories of Thai and the two of English that lie along the timing dimension. The two Spanish categories differ from those of English in that Spanish voiced stops are produced with VOT values that lead the consonant release, while voiceless stops show VOT upon release or immediately thereafter. In the present study, native speakers of Latin American Spanish identified stops synthesized with VOT varying in small steps. Their responses showed a fairly good fit between production and perception, differing from English in the expected direction. Some listeners also discriminated the variants in a psychoacoustic test format. The latter results, along with the earlier ones for English and Thai, suggest that discriminability is largely determined by language experience, although some subjects reveal considerable sensitivity to changes in the acoustic signal at some remove from the phonemic boundary.

Perception of Voice Timing in Spanish Stop Consonants

In earlier acoustic analysis of Spanish, we showed that speakers differentiate voiced /bdg/ from voiceless /ptk/ in word-initial position by means of voice onset time (VOT), the temporal relation between the onset of glottal pulsing and acoustic features of supraglottal articulation. More recently, we demonstrated the perceptual efficacy of VOT for Thai, a language with three categories along the timing dimension, and English, a language with two categories. The two Spanish categories differ from those of English in their placement of the phonetic boundary. Spanish voiced stops are produced with VOT values that lead the consonant release, while voiceless stops show VOT upon release or immediately thereafter. In the present study, native speakers of Spanish identified randomized recordings of stops synthesized with VOT varying in small steps. Labeling responses showed a fairly good fit between production and perception differing from English in the expected direction. The listeners also discriminated the VOT variants in a psychoacoustic test format. The latter results, along with the earlier ones for English and Thai, indicate that discriminability is primarily determined by specific language experience, although general phonetic categories may have some effect.

Coarticulation Effects in the Identification of Final Plosives

In a comparison of the relative contributions of releases and formant transitions to the correct identification of final plosive consonants, Wang [JSHR 2, 66–73 (1959)] found little difference between identification scores of released final /p t k/ and final /p t k/ whose releases had been eliminated by tape-cutting. The former ranged between 90 and 98%, the latter between 73 and 85%. The present investigation compares the perception of released final /p t k/ (following the vowels /i æ a u/) with the perception of final /p t k/ derived from VCV utterances (in which C = /p t k/ and V = /i æ a u/) from which both the second vowel and the consonant release were deleted by tape-cutting. In such VC syllables, the transition to the consonant is modified by the anticipation of the following vowel. For all consonants combined, listeners achieved correct scores of 80.67% for released and 39.00% for unreleased final plosives. The greatest number of misidentifications were associated with /a/. It appears that the anticipation of a following /a/ produces changes in transitions that prevent correct identification of the consonant without the presence of a release. By comparison, initial plosives and plosives derived from VCV utterances by the deletion of the initial vowel were equally easily identifiable: listeners achieved 89.22% correct identification under both conditions.

The location of rhythmic stress beats in English: an experimental study. I.

This study examines the perceived location of rhythmic stress beats in sentences drawn from spontaneous conversation. Subjects listened to each sentence repeatedly from a tape loop and located the beats associated with various syllables in three ways: first, they tapped their fingers in time to the rhythm of the sentence; second, they moved an audible click around in the sentence until they felt it occurred at the same time that they would have tapped in the first task; third, they heard an audible click superimposed on the sentence near a rhythmic beat and were asked to judge whether or not the click "hit the beat". The times of occurrence of the subjects' taps, their placements of the movable click, and their judgments of which clicks hit the beat were compared as beat locations and were found to be in general agreement: the rhythmic beats were closely associated with the onsets of the nuclear vowels of the stressed syllables, but precede those vowel onsets by an amount positively correlated with the length of the initial consonant(s) of the syllable. This finding suggests that rhythmic beats in English are centred around a ballistic release of the initial consonants into the stressed vowels, this release involving respiratory, laryngeal and articulatory activity. The reliability of a subject's beat locating responses to a syllable was positively correlated with the degree of stress on that syllable, i.e., the more stress there is on a syllable, the more accurately can a listener locate the rhythmic beat of that syllable. Both the actual locations of the beats and the correlation of accuracy with degree of stress give meaning to the notion that English is a stress timed language.

The Role of Format Transitions in the Voice-Voiceless Distinction for Stops

Research at the Haskins Laboratories has shown that an acoustic correlate of the distinction between syllable-initial voiced and voiceless stop consonants in English is the onset time of voicing relative to the time of release of the consonant. The present study examines the hypothesis that this critical time on voicing onset is determined by the time when the first formant (and possibly also higher formants) has completed its transition rather than by some fixed absolute time interval of aspiration. A synthesizer was used to generate a series of consonant-vowel stimuli in which (1) the absolute interval between consonantal release and voicing onset, and (2) the duration of the formant transitions, could be manipulated independently. Responses to these stimuli indicate support for the hypothesis. Acoustic analysis of natural consonant-vowel syllables also are consistent with the hypothesis. These findings suggest that the perception of [t] as opposed to [d] in English (for example) is influenced strongly by the presence or absence of a specific acoustic attribute—rapid change in spectrum after voicing onset—and does not require the detection of an absolute voice-onset time from a continuum of values. This attribute may underlie the acquisition of the so-called voice-voiceless distinction by children. [Supported by NIH and by the U. S. Air Force.]

Aerodynamic and Acoustic Events at the Release of Stop and Fricative Consonants

The production of stop and fricative consonants is characterized by (1) rapid mechanical movements at the supraglottal constriction and at the larynx, (2) changes in intraoral pressure and airflow, and (3) a sequence of acoustic events involving turbulence noise generation and glottal exictation of the vocal tract. A theoretical model that accounts for the observed events at the mechanical, aerodynamic, and acoustical levels is developed. This model consists of an acoustic tube with two time-varying constrictions—one at the glottis and one at the point of articulation of the consonant. The airflow and the turbulence noise generated at the constriction are determined from equations of aeroacoustics, and the conditions under which vocal-cord vibration is initiated are established on the basis of theories of vocal-cord operation. The model has been used to make calculations of airflow, pressure, and acoustic output for several classes of voiced, aspirated, and glottalized stop and fricative consonants. The calculations suggest that these consonants fall into natural classes with well-defined and stable acoustic correlates, based on whether the glottis is in a normal, spread, or constricted configuration and whether the vocal cords are stiff or slack. [Supported by NIH and by the U. S. Air Force.]

Comparison of Subglottal Air Pressures Associated with /p/ and /b/

Intraoral air-pressure differences between voiced-voiceless cognates have been attributed by most investigators to a transglottic pressure drop associated with voicing. This assumes pressures below the glottis of essentially the same magnitude during the production of these sounds. The present study tests this assumption. Airflow and subglottal air pressure were recorded as 10 adult males spoke syllables containing /p/ and /b/ in the intervocalic position. Each uttered the syllables stressing the initial vowel and again stressing the final vowel. The airflow trace was used to provide reference points for air-pressure measurement. First, pressure values were measured at a point of no air flow prior to consonant release lip opening and second, the pressure point that corresponded with maximum airflow after release. Analysis of these values showed no significant difference between the pressures at the two points of measurement nor between measures obtained for /p/ and /b/. Differences were found related only to stress. Pressures were greater when the final vowel was stressed.