Stop Vowel Research Articles

Effects of context and speaking rate on liquid–stop sequences: A reassessment of traditional acoustic cues

The allophonic variation of liquids in American English prompted a study examining the findings of Mann [Percept. Psychophys. 28, 407–412 (1980)], Fowler etal . [Percept. Psychophys. 48, 559–570 (1990)], and Liberman [Speech (1996)], who used the F3 transition as the primary cue for synthesizing [da] vs [ga] following [al] and [ar]. The disyllables /al–da/, /al–ga/, /ar–da/, and /ar–ga/ were recorded in a carrier sentence by two male and two female speakers of midwestern English at three different speaking rates, and the spectra and durations of the liquids, stops, and vowels were analyzed. Acoustic analyses indicate structural variation: The F3 transitions of [da] and [ga] by themselves are not stable correlates in these contexts, notwithstanding data used in arguments for the perceptual invariance of articulatory gestures in liquid–stop sequences. Regularities were also found: The liquids systematically affect the F2 and F3 of the vowel in the following stop–vowel sequence, not only the onset value of F3. In addition, each speaker encoded segments in his/her own systematic way over variable rates, which extends Gay’s findings [J. Acoust. Soc. Am. 63, 223–230 (1978)] to tokens with coarticulatory effects. These data will help assess alternative acoustic cues in future perceptual experiments.

Modeling processing dependences in speech perception.

The goal of this work was to determine whether acoustic cues to the phonetic features of a segment are processed independently. For stop–vowel (CV) syllables, one point of controversy concerns whether place information (as conveyed by formant transitions) and voicing information (as conveyed by VOT) are processed mutually independently. In this experiment, a replication of Massaro and Oden [J. Acoust. Soc. Am. 67, 996–1013 (1980)], F2 and F3 transitions and VOT were manipulated as place and voicing cues in CVs. Members of a family of multinomial models, including FLMP, were fitted to each subject’s data. FLMP, which assumes that processing is independent at both the perceptual and decisional levels, failed to account for the data. Two other models retained the perceptual independence assumption, but one assumed the voicing decision was contingent on the outcome of the place decision, and the other assumed the opposite contingency. Both fitted better than FLMP, but nevertheless were rejected on statistical grounds. A model assuming decisional independence, but perceptual dependence, fitted each subject’s data well. It was concluded that the formant transitions and VOT were not perceived independently, but that the voicing and place decisions were made independently. [Work supported by NIDCD (2‐T32‐DC‐00013‐16).]

Articulatory kinematics in VCV sequences

The present study was designed to contribute information on tongue, lip, and jaw movements in sequences of vowel–stop–vowel sounds. The speech material consisted of VCV sequences, with all possible combinations of the vowels /i, a, u/ and the stop consonants /p, t, k/; stress occurred on the second vowel. Receiver coils were attached to the upper lip, lower lip, jaw, tongue tip, and tongue blade; their movements were tracked using an electromagnetic transducing system with three transmitter coils. Results from one speaker suggest the following. When plotted in an occlusal plane reference system, tongue movement trajectories are generally curved. In addition, when the vowels occur in the context of the consonants /t/ and /k/, the tongue trajectories commonly loop back across themselves, thus forming closed loops. Jaw movements showed both curved and linear trajectories for all consonantal contexts. Predictably, tongue movements towards and away from the point of tongue–palate contact showed influences from both the preceding and the following vowel. The temporal and spatial coordination of lip, tongue, and jaw movements in speech production will be discussed. This discussion will focus on the gestures underlying speech motor control and the blending and aggregation of these gestures. [Work supported by NIH.]

Age-related differences in the organization of articulatory gestures for /schwa/♯ stop–vowel sequences.

Acoustic analyses were performed on /schwa/♯ stop–vowel samples from adults and from children (3, 5, and 7 years of age). Stops were {t k d} and vowels were {i ɑ u}. Temporal measures indicated that, compared to adults: (1) children’s /schwa/’s were generally longer and length was more affected by the voicing of the following stop; (2) children’s closure durations were longer; and (3) acoustic voice onset times for 5- and 7-year-olds’ samples were similar to those of adults, but 3-year-olds’ /t/’s were shorter. Measures of formant frequency indicated that: (1) children’s F1’s were similar to those of adults through the entire utterance; (2) children’s F2’s in the stop aspiration noise and at vowel onset showed greater effects of the upcoming vowel than adults; and (3) children’s F2’s at vowel onset showed greater effects of place of closure for the previous stop. Taken together, these results suggest that the organization of articulatory gestures for children between 3 and 7 years of age does not yet show mature phonologic structure. [Work supported by NIH Grant DC-00633.]

On spectral coarticulation in stop–vowel–stop syllables: implications for automatic speech recognition

Stop–vowel–stop syllables of Finnish occurring in a constant frame were studied for coarticulatory influences using machine classifications of auditory spectra spanning the portion of the syllables between the two silent stop occlusions. A simple distance metric was used in the classifications which observed only the total spectrum shape with no further feature extraction. Violating the assumptions of linearity and invariance often endorsed by automatic speech recognition systems, the results witness extensive parallel acoustic transmission of phoneme information. The results also show that, rather than being noise in the signal, the systematic context-dependent variation is a factor that can be used for benefit to increase the amount of phonetic information extractable from the speech signal.

The vowel-dependence of gross spectral cues to place of articulation of stop consonants in CV syllables

It has been argued by Stevens and Blumstein that the acoustic cues of stop consonant place of articulation residing in the gross spectrum, sampled at the release burst onset, are invariant across different vowel contexts. It is shown in this paper that the conclusion of in variance does not follow from the experimental findings of the proponents of the theory of acoustic invariance, and the present results indicate that the vowel exerts a reliable influence on the stop–vowel waveform from the very beginning in terms of gross spectral properties. Also, the consonant influence is seen to be strong throughout the 45 ms interval from the beginning of the CV syllable. Implications of these coarticulatory effects to theories of speech perception are discussed.

Adaptation and contrast in the perception of voicing.

The results of experiments using selective adaptation with stop consonants have been interpreted in terms of auditory feature detector fatigue, phonetic feature detector fatigue, and response contrast. In the present studies, both a selective adaptation procedure and a procedure involving paired comparisons between successively presented stimuli were used to sort out these explanations. A fricative-stop-vowel syllable ([spa]) was constructed using an [s], followed by 75 msec of silence, followed by a 10-msec voice onset time [ba]. The perceived phonetic identity of this syllable was [p] even though the spectral structure of the stop vowel within this syllable was identical to a stimulus from the [ba] end of a [ba]-[pha] test series. As adaptors, the [spa] and [ba] endpoint syllables had identical effects. In paired-comparison procedure, the [spa] caused an ambiguous test item to be labeled "B," whereas the [ba] caused the test item to be labeled "P." Results of these experiments indicate that neither response contrast nor phonetic feature detection are involved in selective adaptation effects found for a voicing stop-consonant series. Results are interpreted as supporting the position that selective adaptation effects arise at an early, auditory level of processing that is responsive to the spectral overlap between adaptor and test items.

Influence of encoding and acoustic similarity on the ear advantage and lag effect in dichotic listening.

Experiments with simultaneous and time lag dichotic listening conditions were used to test two hypotheses concerning the right ear advantage and lag effect in dichotic listening. One hypothesis is based on the similarity of acoustic spectra, and the other is based on a categorization of speech sounds as being either encoded or not encoded. Natural vowels and consonant-vowel syllables were used to obtain seven different types of speech stimuli: stop vowel syllables, fricative vowel syllables, stop burst noise, fricative noise, stop vowel transitions, fricative vowel transitions and steady state vowels. The presentation conditions were monaural, simultaneous dichotic, and dichotic with interaural time delays of 15, 30, 60, and 90 msec. With monaural presentations, all stimuli were identifiable above chance levels. For the simultaneous dichotic condition, significant right ear advantages occurred for stop vowel syllables, fricative vowel syllables, stop burst noise, and steady state vowels. For the time lag conditions, stop vowel syllables, stop bursts, and fricative noise produced consistent lag effects, but steady state vowels produced consistent lead effects. In general, the results gave stronger support to the hypothesis of acoustic similarity than to the encoding hypothesis in that stop burst noise produced both a right ear advantage and a lag effect whereas consonant-vowel transitions produced neither a right ear advantage nor a lag effect.