Ellen M. Parker

Trading relations in speech and nonspeech

Two acoustic variables that correlate with the distinction between intervocalic [b] and [p] are closure duration and presence or absence of low-frequency glottal pulsing during the closure interval. These variables may be considered to exhibit a trading relation (Repp, 1982), to the extent that a longer closure is required to perceive the consonant as voiceless when glottal pulsing is present than when it is not. Such trading relations have been interpreted as reflecting a special speech mode of perception. In the present experiments, we demonstrated a trading relation between closure duration and closure pulsing for a set of [aba]-[apa] stimuli. Next we showed that a similar effect could be obtained with square-wave analogue stimuli that mimicked the segment durations and peak amplitudes of the speech stimuli but that were not phonetically categorizable. This nonspeech trading relation depended on the degree of spectral continuity between the low-frequency pulsing and the adjacent portions of the square wave. The implications of these results for the speech mode hypothesis are discussed.

Auditory constraints on the perception of voice-onset time: the influence of lower tone frequency on judgments of tone-onset simultaneity

The experiments reported employed nonspeech analogs of speech stimuli to examine the perceptual interaction between first-formant onset frequency and voice-onset time, acoustic cues to the voicing distinction in English initial stop consonants. The nonspeech stimuli comprised two pure tones varying in relative onset time, and listeners were asked to judge the simultaneity of tone onsets. These judgments were affected by the frequency of the lower tone in a manner that parallels the influence of first-formant onset frequency on voicing judgments. This effect was shown to occur regardless of prior learning and to be systematic over a wide range of lower tone frequencies including frequencies beyond the range of possible first-formant frequencies of speech, suggesting that the effect in speech is not attributable to (tacit) knowledge of production constraints, as some current theories suggest.

A further parallel between selective adaptation and contrast.

It is generally believed that selective adaptation effects in speech perception are due to a reduction in sensitivity of auditory feature detectors. Recent evidence suggest that these effects may derive instead from contrast. In a further test of the contrast hypothesis, we conducted two experiments each involving both adaptation and contrast sessions with matching stimulus sets. During the adaptation sessions of Experiment 1, subjects identified two series of velar stimuli varying in voice onset time, [ga]-[kha] and [gi]-[khi], before and after adaptation with of the following stimuli: [ga], [kha], [gi], and [khi]. In the contrast session, subjects identified either of two ambiguous test items (drawn from near the phonetic boundaries of the [ga]-[kha] and the [gi]-[khi] series) following a single presentation of [ga], [kha], [gi], or [khi]. For both the adaptation and contrast sessions, (a) the [--a] test items were more greatly affected (in a contrast direction) by the [--a] than by the [--i] adaptor/context stimuli, and (b) the [--i] test items were not differentially affected by the [--1] and [--i] adaptor/context stimuli. An analogous design was used in Experiment 2, except that the stimulus sets varied in pitch rather than vowel quality. For both the adaptation and contrast sessions, the test items were not differentially affected by the pitch of the adaptor/context stimulus. These parallel results provide further evidence that adaptation effects are actually a form of contrast.

A further parallel between selective adaptation and response contrast

It is generally believed that selective adaptation effects in speech perception are due to a reduction in sensitivity of auditory feature detectors. We have recently presented evidence that these effects may derive instead from response contrast. In a further test of the contrast hypothesis, we conducted both an adaptation and a contrast experiment with matching sets of stimuli. In the adaptation experiment, subjects identified two series of velar stimuli varying in VOT, [ga‐kha] and [gi‐khi], before and after adaptation with each of the following test stimuli: [ga], [kha], [gi], and [khi]. In the contrast experiment, subjects identified either of two ambiguous test items (drawn from near the phonetic boundaries of the [ga‐kha] and the [gi‐khi] series) following a single presentation of [ga], [kha], [gi], or [khi]. For both the adaptation and the contrast experiments, (1) the [Ca] test items were more greatly affected (in a contrast direction) by [Ca] than by the [Ci] adaptor/context stimuli, and (2) the ...

Resonance versus source characteristics in perceiving spectral continuity between vowels and consonants

Using nonspeech analogs of VCV sequences modeled as a gap between square waves, Parker et al. [Percept. Psychophys. 39, 129–142 (1986)] showed that pulsing in the gap makes the gap sound shorter. This interaction may underlie the common covariation between closure voicing and closure duration in intervocalic stops. However, the gap duration judgments only depended on pulsing when the analogs F0 fell into and rose out of the gap, implying that to influence perceived gap duration, pulsing must be spectrally continuous with the flanking sounds. Extending this result to speech stimuli is difficult because both F1 and F0 fall into and rise out of voiced stop closures. In the present experiment, the contribution to spectral continuity of independently varying constant, falling‐rising and rising‐falling F1 and F0 patterns are compared with single formant vowel analogs. These nine spectral patterns flanked gaps varying from 20–120 ms in 10‐ms steps. Pulsing produced significantly more short gap judgments when F1...

Identifying vowels in CVC syllables: Effects of inserting silence and noise

Listeners were asked to identify natural vowels in /d_d/ context under various deletion conditions. Deletion intervals, which were centered about the syllable midpoint, ranged from 60% to 90% of the syllable duration and contained either silence or broadband noise. In one condition, the three syllable types were /did/, /ded/, and /dud/; in the other condition, the three syllable types were /did/, /ded/, and/dAdV. Identification performance in the 60% and 70% deletion conditions was not substantially worse than for full syllables. Even the 90% deletion conditions yielded performance well above chance, indicating that significant coarticulatory information for vowels is contained in the first and last 10 or 15 msec of the syllable. For the short-vowel stimuli (/did/, /d∈d/, and /dλV) in the 90% deletion condition, a series of discriminant analyses were performed to assess the relative contribution of several acoustic variables to the statistical separation of the vowel categories. Several different combinations of acoustic variables (including formant frequencies at particular temporal locations and formant frequency differences over time) were sufficient to yield significant separation of the three vowel categories. However, in general, the performance of the discriminant classification program correlated only weakly with the identification performance of the listeners.

Trading relations in nonspeech

Two acoustic features shown to be sufficient in specifying the distinction between intervocalic /b/ and /p/ are closure duration and the presence or absence of a low‐frequency voicing pulse during the closure interval. These features of the signal can be considered to exhibit a trading relation [B. Repp, Psych. Bull. 92, 81–110 (1982)] to the extent that a longer closure is required to perceive the consonant as voiceless when the pulse is present than when it is not. This study addressed the issue of whether a trading relation of the type found in speech can be demonstrated for nonspeech as well. Two types of square wave analogs, modeled after a speech series ranging from “rabid” to “rapid,” were constructed. One set of nonspeech stimuli contained frequency transitions into and out of the silent interval while frequency for the other nonspeech stimuli remained constant. Preliminary (AXB) identification results indicate that in the speech condition, the presence of the pulse affects a shift in the ID bound...