Sharon Y. Manuel

Discovering phonetic coherence in acoustic patterns

Despite spectral and temporal discontinuities in the speech signal, listeners normally report coherent phonetic patterns corresponding to the phonemes of a language that they know. What is the basis for the internal coherence of phonetic segments? According to one account, listeners achieve coherence by extracting and integrating discrete cues; according to another, coherence arises automatically from general principles of auditory form perception; according to a third, listeners perceive speech patterns as coherent because they are the acoustic consequences of coordinated articulatory gestures in a familiar language. We tested these accounts in three experiments by training listeners to hear a continuum of three-tone, modulated sine wave patterns, modeled after a minimal pair contrast between three-formant synthetic speech syllables, either as distorted speech signals carrying a phonetic contrast (speech listeners) or as distorted musical chords carrying a nonspeech auditory contrast (music listeners). The music listeners could neither integrate the sine wave patterns nor perceive their auditory coherence to arrive at consistent, categorical percepts, whereas the speech listeners judged the patterns as speech almost as reliably as the synthetic syllables on which they were modeled. The outcome is consistent with the hypothesis that listeners perceive the phonetic coherence of a speech signal by recognizing acoustic patterns that reflect the coordinated articulatory gestures from which they arose.

Acoustic properties of /h/

The aim of this paper is twofold: (1) to investigate the physical mechanisms of sound generation for the consonant /h/ and (2) to examine the timing of supraglottal and glottal movements of /h/. Utterances in which /h/ was present or absent (e.g., “new heart” versus “new art”) were analyzed acoustically and contrasted. The corpus consisted of about 20 such utterances repeated several times by three speakers. The acoustic data showed evidence of breathy voicing at the /h/‐vowel boundary in all cases, and that generation of turbulence noise during the consonant occurred both in the vicinity of the glottis (aspiration noise) and the vicinity of the supraglottal constriction (frication noise). The relative contribution of the two noise sources depended on the vowel, with greater frication noise occurring for high vowels. When an /h/ was in position between two vowels or glides, it generally added little or no duration to the utterance, relative to the contrasting utterance with no /h/. Implications for the ph...

Oral and glottal gestures and acoustics of underlying /t/ in English

In English, an underlying /t/ can be produced in various ways. The glottal gestures may involve glottal attack (glottal stop), spreading, or voicing. The oral gestures may include complete or incomplete alveolar closure; or, in the case of place assimilation (e.g., “notebook” surfacing as [nowpbUk]), they may entail another oral gesture. Thus some surface variation is due to the particular gestures used. However, since the relative timing of glottal and oral gestures may effectively preclude acoustic realization of some of them, the acoustic record alone may not reveal just which gestures have actually been produced, and in what order. Therefore, some perceived variation may be due to differences in timing and relative strength of articulatory gestures, rather than changes in the set of gestures. This investigation of alveolar stops of two talkers focuses in particular on the occurrence of glottalized /t/ and shows differences in articulatory patterning depending on talker, speaking style, and phonetic context. Data are based on electropalatography, transglottal illumination, fiberoptic video of the larynx, oral air pressure, and acoustic measures. [Work supported by NINCDS grants to MIT and to Haskins Laboratories.]

Vowel reduction and perceptual recovery in casual speech.

A pervasive characteristic of casual (normal conversation) English is the apparent deletion of unstressed vowels like the first vowel in the word ‘‘support.’’ One might suppose that if the first vowel in ‘‘support’’ were deleted, ‘‘support’’ would become homophonous with ‘‘sport.’’ Acoustic and physiological data are reported which suggest that in fact when speakers appear to be deleting an unstressed schwa, they are often actually omitting only the oral gestures for the vowel. The glottal gestures stay much as they are in careful speech and are tied to the remaining oral gestures much as they are in careful speech. Also reported are perceptual data which suggest that listeners are sensitive to the acoustic consequences of these residual patterns, and can use this information to distinguish between ‘‘sport’’ and reduced versions of ‘‘support’’ [see also J. Fokes and Z. S. Bond, J. Acoust. Soc. Am. Suppl. 1 85, S85 (1989); J. Fokes and Z. Bond, Proc. XII Int. Congr. Phon. Sci. 4, 58–61 (1991)]. These resul...

Output constraints and cross‐language differences in coarticulation

Coarticulatory patterns have been shown to vary from language to language, from speaker to speaker, and from one style or rate of speaking to the next. Many of these differences might be explained by assuming (1) that phones are associated with ranges of acceptability, that is, “acoustic output constraints,” which must not be violated by coarticulation‐induced variability and (2) that these output constraints vary from language to language, speaker to speaker, etc. For this notion to have any predictive value, it is necessary, in turn, to be able to predict, to some extent, just what the acoustic output constraints are, and how they might vary. It might be expected, for example, that output constraints relax in casual speaking styles. The paper is particularly concerned with output constraints that can be understood in terms of systems of phonemic contrast. Specifically, it is predicted that when the phonemes of a language are widely spread out in the acoustic space, then the acoustic output constraints o...

Place of articulation for consonants: Comparing nasals and stops, and syllable position

This study reports on measurements of transitions of the second formant (F2) for syllable‐initial and syllable‐final nasal and stop consonants in English, as cues to place of articulation. F2 values and the amount and direction of F2 changes in a 20 ms interval were determined in vowels adjacent to consonant ‘‘implosion’’ and release. The corpus included consonants in sentences and isolated nonsense syllables. The F2 transitions for a given place of articulation are roughly similar for nasals and stops for different syllable positions, as expected, but some interesting differences exist. Since the stop‐consonant bursts influence the spectrum sampling point relative to release, there are some shifts in F2 onset frequencies for stops, as compared to nasals. Syllable‐initial alveolars show additional manner differences, which can be attributed to differences in tongue configurations for alveolar nasals relative to stops in back‐vowel contexts. These latter shifts are tentatively ascribed to differential coar...

Correlating movement and acoustic measures of nasalization

Previous research using physiological instrumentation (e.g., Velotrace) has shown systematic effects of syllable structure and stress on velum movements [Krakow, in Phonetics and Phonology V (Nasals, Nasalization, and the Velum, edited by Huffmann and Krakow (Academic, San Diego, 1993), pp. 87–113]. However, what has not been investigated previously is the correspondence between the time‐varying position of the velum in such data and the timing and magnitude of acoustically detectable nasalization. A new technique developed by Chen [ 3283 (A) (1994)] for identifying the spectral effects of nasal coupling was found to provide a valuable way of linking the acoustic and articulatory changes related to velopharyngeal port aperture. This paper reports on acoustic‐to‐articulatory mapping using acoustic data collected in concert with Velotrace data for a number of different utterance types, including those with stressed versus unstressed syllables and those with syllable‐initial versus syllable‐final nasals. One...

Speakers nasalize /edh/ if it is preceded by /n/, but listeners don’t care—They still hear /edh/

Presumably as a result of coarticulation, /edh/ often assimilates to a preceding /n/ in phrases like ‘‘win those,’’ but this assimilation is not complete for all features. With respect to the feature [nasal], the assimilation is often radical. The entire consonant region in the middle of the two‐word sequence is nasalized. However, acoustic evidence suggests that contextually nasalized /edh/ retains its dental place of articulation. Specifically, F2 is considerably lower at the release of a contextually nasalized /edh/ than at the release of a true /n/, as would be expected for a dental consonant. Perception tests show that listeners can generally tell the difference between natural tokens of pairs like ‘‘win nos’’ and ‘‘win those,’’ even when the /edh/ is completely nasalized. In addition, a synthetic stimulus continuum was constructed in which items differed only with respect to F2 frequency in the vicinity of the nasal consonant regions of phrases like ‘‘win nos.’’ Listeners systematically reported hea...

Perceptual effects of attention to phonetic versus “musical” properties of sinewave speech

A continuing debate in speech perception research has been whether phenomena such as categorical perception and trading relations are explained by basic psychoacoustic mechanisms or rather by a specialized phonetic mode of perception. Sinewave speech analogs have provided an opportunity to test these alternatives, since the same stimuli can be heard as either distorted speech or nonspeech sounds by different groups of listeners. The psychoacoustic hypothesis would predict similar between‐group performance, while the phonetic hypothesis would predict systematic group differences. Several recent findings have favored a distinct phonetic mode in perception of sinewave speech. Best and Studdert‐Kennedy [Proc. 10th Int. Congr. Phon. Sci (1983)] used a labeling test with a sinewave‐analog /ra/‐/la/ continuum, and found that speech‐biased listeners performed in a clearly categorical manner, whereas “music”‐biased listeners performed near chance. The present research strengthened that finding, by employing AXB id...