• Volume/Page
  • Keyword
  • DOI
  • Citation
  • Advanced
   
 
 
 

Journal of the Acoustical Society of America

SECTION LISTING

BROWSE VOLUMES

Year Range: 
Search Issue | RSS Feeds RSS
Previous Issue Next Issue

Nov 1980

Volume 68, Issue S1, pp. S1-S116

Return to All Sections
back to top
RSS Feeds
back to top Session E. Speech Communication I: Speech Perception I
Contributed Papers
FREE

Effects of speaking rate on vowel identification (A)

Thomas L. Johnson and Winifred Strange

J. Acoust. Soc. Am. Volume 68, Issue S1, pp. S9-S9 (1980); (1 page)

Online Publication Date: 11 Aug 2005

Full Text: | Download PDF

Show Abstract
To examine the role of context in providing temporal information that supports identification of vowels, eleven t‐vowel‐t syllables were excised from a carrier sentence spoken at both normal and rapid speaking rates by an adult male talker. Separate groups of subjects identified vowels in the syllables presented in the following contexts: (1) in isolation, (2) embedded in the inappropriate rate carrier sentence, (3) embedded in the appropriate rate carrier sentence, and (4) accompanied by the immediately preceding word (“the”) from the rapidly‐produced carrier (minimal context). Vowel identification was highly accurate for normal rate targets in all contexts. For rapidly produced syllables were most frequent for syllables in isolation, intermediate for syllables in inappropriate rate and minimal contexts, and least frequent for syllables in appropriate rate context. Most errors in the identification of rapid targets occurred on “intrinsically long” intended vowels. The results indicate that vowels may be incompletely specified in a single syllable of rapid speech, and that the surrounding speech context may provide temporal information necessary for accurate vowel identification. [Work supported by NIMH, NICHHD, NSF.]
FREE

The effect of language familiarity on vowel discrimination (A)

Batya Levine Jacob, Peter J. Alfonso, and Antonia B. Maxon

J. Acoust. Soc. Am. Volume 68, Issue S1, pp. S9-S9 (1980); (1 page)

Online Publication Date: 11 Aug 2005

Full Text: | Download PDF

Show Abstract
Recent studies have shown that phonetic and phonological context influences vowel identification. For instance, vowels are better identified when presented in a syllabic environment than when presented in isolation [Strange, Verbrugge, Shankweiler, and Edman, J. Acoust. Soc. Am. 60, 213–224 (1976)], and when the syllabic environment is represented by natural speech formant transitions rather than “flipped formants” [Fowler and Shankweiler, J. Acoust. Soc. Am. Suppl. 1 63, S4 (1978)]. The identification of steady‐state vowels presented in concatenated sequences is also influenced by context factors [Alfonso and Kerivan, J. Acoust. Soc. Am. Suppl. 1 66, S86 (1979)]. Finally, familiarity of contextual language‐specific factors affect vowel identification [Gottfried, J. Acoust. Soc. Am. Suppl. 1 65, S6 (1979); Rudnicky, .I. Acoust. Soc. Am. Suppl. 1 65, S110 (1979)]. The study reported here demonstrates that vowel discrimination is also context dependent. Two groups of subjects, native American‐English and native Hebrew speakers, discriminated among six vowels, three that are familiar to both groups, and three that are represented in English but not Hebrew. AX discrimination CVC pairs consisted of the following: two vowels familiar to the Hebrew group, designated (FF), two unfamiliar vowels (UU), or combinations of unfamiliar and familiar vowels (UF). The Hebrew group demonstrated significantly poorer discrimination than the English group on the (UF) pairs. The results and the possible use of unfamiliar speech stimuli in the clinical management of the hearing impaired will be discussed.
FREE

Perception of the oral‐nasal distinction in vowels: A cross‐language study (A)

Patrice Speeter Beddor

J. Acoust. Soc. Am. Volume 68, Issue S1, pp. S9-S10 (1980); (2 pages)

Online Publication Date: 11 Aug 2005

Full Text: | Download PDF

Show Abstract
Cross‐language research with isolated vowels has shown that accuracy of vowel discrimination, unlike consonant discrimination, is not affected by linguistic experience [K. N. Stevens, A.M. Liberman, M. Studdert‐Kennedy, and S. E. G. Öhman, Lang. Speech 12, 1–23 (1969)]. The present study extends cross‐linguistic investigation of vowel perception to the oral‐nasal contrast. Speakers of French, Hindi, and English were tested on identification and discrimination tasks using several oral‐nasal vowel series to determine the effect of phonemic status of nasal vowels on accuracy of discrimination. (The contrast between oral and nasal vowels is phonemic in French and Hindi, but allophonic in English.) Stimulus materials were generated by articulatory synthesis and involved three 11‐step /ba/‐/bā/ series in which size of velopharyngeal port opening increased in equal increments from the oral to nasal endpoints. Results of three‐step oddity discrimination tests for a series incorporating 2.4 sq mm increments showed similar performance for all three language groups. However, performance was generally poor for all subjects. The effect on discrimination accuracy of, two modifications in the series—increased discrimination step size and increased port aperture at the nasal end of the series—was investigated. [Work supported by NICHD, NIMH.]
FREE

Identification of stops and vowels in truncated CV syllables (A)

David B. Pisoni and Diane Kewley‐Port

J. Acoust. Soc. Am. Volume 68, Issue S1, pp. S10-S10 (1980); (1 page)

Online Publication Date: 11 Aug 2005

Full Text: | Download PDF

Show Abstract
How is the information specifying place of articulation and vowel color encoded in a stop‐vowel syllable? Are place cues located in just the burst, or the first 20 ms of a CV waveform as some investigators have suggested? Alternatively, are place cues encoded simultaneously (i.e., in parallel) with vowel information in the formant transitions? To answer these questions, an identification experiment was conducted using the consonants /b, d, g/ paired with the vowels /i, e, a, o, u/ produced by two male talkers. Using computer techniques, five truncated stops were edited from each original syllable. Cuts were made after the release burst, and then the burst plus 1, 3, 5, or 7 pitch pulses. On different days, naive subjects identified either the consonant or the vowel. Identification functions for /b/ and /d/ reached 95% correct within the first 20 ms. Vowels, except for /i/, were identified near chance for the shortest stimuli. The consonant /g/ needed a longer duration, 40 to 50 ms, for correct place identification, although /gi/ was never identified accurately even at the longest durations. We conclude that sufficient acoustic information for identifying the stops /b/ and /d/ can be found within the first 20 ms of a CV syllable, although information for the /g/ extends over somewhat longer durations. Formant transitions appear to carry information about the vowel although somewhat less information about the consonant, except for /g/ before front vowels. [Supported by NINCDS Grant NS‐12179.]
FREE

Rate of acoustic formant transition change on categorical perception (A)

Joyce Schwartz and Paula Tallal

J. Acoust. Soc. Am. Volume 68, Issue S1, pp. S10-S10 (1980); (1 page)

Online Publication Date: 11 Aug 2005

Full Text: | Download PDF

Show Abstract
A comparison was made of the effect of rate of acoustic formant transition change on the categorical perception functions for stop‐consonant‐vowel (CV) syllables with 40‐ and 80‐ms formant transitions. Two place of articulation continua representing the phonemic categories /ba/ and /da/ containing transitions of 40 and 80 ms were presented to subjects in both identification and discrimination tasks. In the 80‐ms signals, the entire formant transition spectrum was extended. Individual identification functions were fit using probit analysis. Extending the duration of formant transitions while keeping the manner of articulation constant significantly effected the identification function. Discrimination for within‐category stimulus comparison was enhanced for the extended formant transitions.
FREE

Normalization of /w‐r/ category boundary variations due to vowel, formant and talker variations (A)

Peter J. Benson and Donald J. Sharf

J. Acoust. Soc. Am. Volume 68, Issue S1, pp. S10-S10 (1980); (1 page)

Online Publication Date: 11 Aug 2005

Full Text: | Download PDF

Show Abstract
Different synthetic speech continua representing different talkers will not necessarily have the same category boundary (CB) even if the continua vary along similar acoustic dimensions. The scales of the continua can be normalized so that the CB's are the same by finding an appropriate normalization function. Since there are an infinite set of such functions, any particular normalization is interesting only if it has a basis in perception and predicts variations in the CB in other continua. Two synthetic /w‐r/ continua were synthesized by varying the separation of F2 and F3. Synthesis parameters were chosen to mimic an adult and a child voice. A third child‐like continuum was created by scaling the adult parameters up by the factor 1.5. The CBs were equilibrated by replotting labeling results against parameter values scaled by (mel F3 – mel F2)/mel F1. This function employs the primary acoustic difference and an indirect measure of vocal tract length (F1), transformed by an approximation to the psychophysical function for pitch perception. Results will be presented that evaluate this and related functions for their value in equating CB for different vowels and talkers and for variations in F1.
FREE

Effects of stimulus intensity on perception of a place cue for stop consonants (A)

Dianne J. Van Tasell, Elizabeth S. Anderson, Todd D. Ericksen, and Lawrence T. Hagen

J. Acoust. Soc. Am. Volume 68, Issue S1, pp. S10-S10 (1980); (1 page)

Online Publication Date: 11 Aug 2005

Full Text: | Download PDF

Show Abstract
Synthetic syllables from a /da/‐/ga/ continuum varying in onset frequency of the third formant transition were presented to normal‐hearing listeners at 60 dB SPL and at 100 dB SPL. The listeners' identification responses showed that syllables were classified as members of a phonetic category less consistently at 100 dB than they were at 60 dB. Discrimination performance then was measured at two different locations on the synthetic syllable continuum, using a same‐different paradigm with correct‐answer feedback. When the standard stimulus was at a location on the stimulus continuum corresponding to maximum differences in identification performance between 60 and 100 dB presentation levels, the measured jnd for third formant onset frequency was substantially larger at 100 dB than it was at 60 dB. Discrimination performance at the two levels was more similar when the standard stimulus was a syllable that had been identified consistently at both 60 and 100 dB. [Work supported by PHS Grant NS‐12125.]
FREE

More adaptation of speech and nonspeech (A)

Donna Kat and Arthur G. Samuel

J. Acoust. Soc. Am. Volume 68, Issue S1, pp. S10-S10 (1980); (1 page)

Online Publication Date: 11 Aug 2005

Full Text: | Download PDF

Show Abstract
In a previous study [A. G. Samuel and E. L. Newport, J. Exp. Psychol.: Hum. Perc. Perf. 5(3), 563–578 (1979)], nonspeech adaptors (440 Hz sawtooth waves and white noise segments, with 0‐ or 80‐ms rise times) caused significant labeling shifts on speech continua. In particular, the abruptly onsetting sawtooth reduced the number of [ba]'s reported (versus [wa]), and the gradually onsetting noise segment reduced responses of [sha] (versus [cha]. These results suggested that (1) speech and nonspeech are processed by the same mechanisms, and (2) these mechanisms are sensitive to the general spectral quality of the input, i.e. whether the spectrum is generally flat (aperiodic) or peaked (periodic). In the present study, the same four adaptors were used on a [dɜɑ]—[ɜɑ] continuum. These stimuli have spectra quite similar to [+ʃɑ]—[ʃɑ], but are voiced, and thus have a low‐frequency periodic component. As with [cha]—[sha], only the gradually onsetting noise adaptor was effective. These results indicate that as suggested in the earlier study, overall spectral quality (“flat” versus “peaked”) is the critical determinant, not the presence or absence of glottal pulsing.
FREE

Infants' discrimination of full and partial cues to place of articulation in stop consonants (A)

Richard N. Aslin and Amanda C. Walley

J. Acoust. Soc. Am. Volume 68, Issue S1, pp. S10-S11 (1980); (2 pages)

Online Publication Date: 11 Aug 2005

Full Text: | Download PDF

Show Abstract
Stevens and Blumstein [J. Acoust. Soc. Am. 64, 1358–1368 (1978)] have proposed that the shape of the onset spectrum provides contextually invariant information about place of stop consonant articulation for CV syllables and that these primary spectral attributes underlie the. infant's ability to discriminate place differences. According to this view, contextually variable formant transitions constitute only secondary or “learned” cues to place of articulation. Prelinguistic infants are, therefore, assumed to be incapable of discriminating place differences in two‐formant stimuli which supposedly lack invariant spectral attributes. Our analysis revealed, however, that the two‐formant labial and velar CV stimuli are spectrally very similar to their full‐formant counterparts. Therefore, if spectral shape does cue place of articulation, infants should actually discriminate these stimuli. An operant head‐turning paradigm was employed to test this hypothesis with 6–9 month old infants. The results indicated that infants can discriminate two‐formant stimuli—a finding which renders the distinction between primary versus secondary cues invalid. Moreover, because several infants discriminated the two‐formant alveolar and velar stimuli, whose onset spectra are very similar, it is probably not the shape of the spectrum at stimulus onset which mediates infants' discrimination of place of articulation in stop consonants. [Work supported by NICHHD, NIMH, and Research Council of Canada.]
FREE

Speech discrimination by macaques: enhanced discrimination at the phonetic boundaries between speech‐sound categories (A)

Patricia K. Kuhl and Denise M. Padden

J. Acoust. Soc. Am. Volume 68, Issue S1, pp. S11-S11 (1980); (1 page)

Online Publication Date: 11 Aug 2005

Full Text: | Download PDF

Show Abstract
Discrimination of speech‐sound pairs drawn from a computer‐generated continuum in which syllables varied along the Place of Articulation phonetic feature (/bae, dae, gae/) was tested with macaques. The acoustic feature that was varied along the 15‐step continuum was the starting frequencies of the first and second formant transitions. Discrimination of stimulus pairs separated by two steps was tested along the entire continuum in a same‐different task. Results demonstrated that peaks in the discrimination functions occur for macaque listeners at the “phonetic boundaries” which separate the /b‐d/ and /d‐g/ phonetic categories for human listeners. This effect of greater discriminability at the phonetic boundaries in macaque listeners replicates that seen in human listeners as young as one month of age. The data demonstrate that the results previously obtained with nonhuman listeners on the voiced‐voiceless (/b‐p/, /d‐t/, /g‐k/) phonetic feature can be generalized to another phonetic feature; this finding is a critical element in the general acceptance of the notion that speech‐sound contrasts were chosen to be highly distinctive to the mammalian auditory system. [Supported by NSF.]
FREE

Phonetic boundaries and judgmental central tendency (A)

Blas Espinoza‐Varas

J. Acoust. Soc. Am. Volume 68, Issue S1, pp. S11-S11 (1980); (1 page)

Online Publication Date: 11 Aug 2005

Full Text: | Download PDF

Show Abstract
What determines that phonetic boundaries be placed at some specific loci of speech stimulus continua? According to one view, boundaries would be determined by relatively stable, built‐in properties such as psychophysical thresholds, selectively tuned feature detectors, or articulatory constrains. Alternative views, e.g., central tendency in psychological judgments [R. S. Woodworth, Experimental Psychology (Holt, New York, 1938), pp. 445], and adaptation‐level theory [H. Helson, Psych. Rev. 55, 297–313 (1948)], suggest that category boundaries are learned psychological standards that subjects develop when classifying stimuli along a dimension. These psychological standards tend to match some central tendency measure of the stimulus distribution. This paper re‐evaluates the adequacy of these alternative views with existing labelling data for speech continua of VOT, duration of stop closure, and onset and duration of formant transitions. For most of these data, the observed values of the phonetic boundaries approximated well the arithmetic mean of the stimulus distribution. In addition, changes in the phonetic boundary appeared to follow changes in the arithmetic mean that resulted from varying the range or shape of the stimulus distribution. The contribution of judgmental central tendency to the development of phonetic boundaries will be discussed. [Supported in part by NRC Postdoctoral Fellowship to the author.]
Return to All Sections
Close

Home Publications Meetings Contact ASA Site Map Back to Top

Copyright © Acoustical Society of America

close