Auditory memory and vowel anchoring in CVC contextsJ... Schwab Department of Psychology, SUNY/Buffalo, 4230 Ridge Lea Rd., Buffalo, NY 14226 Previous vowel anchoring experiments have s
Trang 1Auditory memory and vowel anchoring in CVC contexts
J R Sawusch , H C Nusbaum , and E C Schwab
Citation: The Journal of the Acoustical Society of America 65, S113 (1979); doi: 10.1121/1.2016954
View online: https://doi.org/10.1121/1.2016954
View Table of Contents: https://asa.scitation.org/toc/jas/65/S1
Published by the Acoustical Society of America
Trang 2collected from three male and three female English speakers Three
groups of 12 subjects distinguished the smiled token from its straight-
faced counterpart in forced-choice identification tasks: Group 1 was
instructed to choose the smiled sound, Group 2 the happier, and
Group 3 the sadder Groups 1 and 2 selected the smiled utterance
significantly more often than chance for all six speakers Group 3
chose the straight-faced utterance as consistently, for four speakers,
but for one speaker reliably picked the smiled one The results
suggested that particular cue combinations may be identified as
smiling specifically, whereas others may contribute to perception of
emotionally in general Acoustic analyses showed that smiling raised
the second and third formant frequencies for all speakers, and
fundamental frequency, amplitude, and/or duration for some
Duration and frequency both increased substantially during smiling
only for the speaker whose smiled utterances were also judged as sad
SS10 Auditory memory and vowel anchoring in CVC contexts
J R Sawusch, H C Nusbaum, and E C Schwab (Department of
Psychology, SUNY/Buffalo, 4230 Ridge Lea Rd., Buffalo, NY 14226)
Previous vowel anchoring experiments have shown some evidence
of multiple processes in the contrast effects found for vowels In the
present experiments vowels were embedded in CVC context
([CiC] to [CIC]) and presented to subjects in ABX format as well as
standard identification in both equiprobable control and anchoring
conditions For the [CIC] end of our continua, the degree of change
in category boundaries due to anchoring was found to vary with the
degree to which the stimuli were perceived categories (from the
ABX data) However, for the [CiC] end of our continua, the
degree of anchoring was nearly constant and independent of the
consonantal context and the degree of categoricalness in the ABX
data These data offer further support for our interpretation of vowel
anchoring and adaptation results in terms of two distinct
perceptual processes with one of these tied changes in auditory
memory [Work supported by NSF and NIMH.]
SSll Identification and discrimination of a new linguistic contrast
R N Aslin, D B Pisoni, B L Hennessy, and A J Perey
(Psychology Department, Indiana University, Bloomington, IN 47405)
For many years there has been a consensus among investigators
that the linguistic experience of an individual exerts a profound
and quite often permanent effect on the perceptual abilities required
to identify and discriminate speech sounds Based on the results of
a small number of studies, it has been further assumed that selective
modification of phonetic perception cannot be accomplished easily
and quickly in the laboratory with simple discrimination training
techniques involving only a few hours of practice These findings
have been interpreted as support for the notion that the underlying
neural mechanisms become so finely aligned by the linguistic
environment that they can no longer be modified or "returned" in
adults In the present paper we report the results of three
experiments that systematically examined the acquisition of a new
linguistic contrast in voicing Laboratory training procedures were
implemented with a small computer in a real-time environment to
establish a three-way contrast among voiced, voiceless unaspirated,
and voiceless aspirated stops New perceptual categories emerged
quite rapidly for most monolingual English subjects after only a few
minutes and subsequent perceptual tests revealed reliable and
consistent labeling and discrimination functions for all three
categories These new results differ quite substantially from all
previous studies in demonstrating that the adult perceptual system is,
in fact, quite plastic and can be modified selectively by environmental
input [Supported by NIMH and NINCDS.]
SS12 Abstract withdrawn
SS13 Influence of task demands on the categorical versus continuous
perception of vowels N Cowan and P A Morse (Department of Psychology, University of Wisconsin, Madison, WI 53706) Various task constraints were imposed on subjects making same/ different, AX discriminations within and between phonetic categories The experiments employed four 50-ms, synthetic vowels from an/i/-/I/continuum and five interstimulus delays (0-2000 ms)
In the first experiment, stimuli were presented with all trial types randomized together or with between- and within-category shifts in separate trialblocks Between-category performance was superior with separate blocks, whereas within-category performance did not differ in the two procedures A second experiment involved AX presentations to subjects who were or were not initially informed
of the intended category labels Surprisingly, category knowledge improved within-category discriminations In a third experiment, subjects were required to identify both vowels in an AX pair Between-category discrimination was higher, and within-category discrimination lower, with AX identification as compared to a same/different response The results of these experiments suggest that a stable phonetic and an unstable auditory short-term memory both contribute to within-and between-category discrimination, with some labeling occurring on a comparative basis Furthermore, task demands alter the relative contributions of auditory and phonetic
memory
SS14 Identification of temporal order in vowel sequences J E Kerivan (NSMRL, Auditory Department, Submarine Base, Groton,
CT 06340) P J Alfonso and S P Bornstein (Department of Speech, University of Connecticut, Storrs, CT 06268)
The effects of element speed and element pitch contour were studied for synthetic vowels /i, a•, a, u/ in various permuted sequences Four trained subjects were asked to identify temporal order in this closed response vowel set A high-level stimulus uncertainty procedure was used Subjects were near perfect performance at element durations ranging from 90-150 ms They were above d' = 1.0 level at 30 ms per element and d' = 2.0 at 60 ms per element Performance on individual sequence permutations varied little at 90-150 ms per element but varied between chance and slightly above d' = 2.0 at both 30 and 60 ms per element The effect of normalizing element pitch contours was negligible for individual sequence permutations at all element speeds The duration threshold of vowel sequence identifiability (125- 200 ms per element) reported by other investigators using a recycling presentation format is considerably longer than these 30-60 ms per element thresholds A more sensitive approach to determine absolute duration thresholds for vowel sequences appears to be with trained subjects using a single sequence versus a recycling sequence
method
Sl13 J Acoust Soc Am., Vol 65, Suppl No 1, Spring 1979 50th Anniversary Meeting Sl13