Perception of voicing in English affricates and fricatives

Abstract

Two series of experiments were conducted to study the effects of consonant and vowel duration on the perception of the voiced–voiceless distinction for the English affricates [?] vs [?] and the fricatives [v] vs [f] and [z] vs [s]. The voiceless members of each pair of speech sounds were recorded by a male speaker in a CV environment with the vowel [a] ([?a], [fa}, and [sa]) and then duplicated a number of times. The frication of each syllable was then shortened in small steps by removing frication from just prior to the vowel and then closing the gap. Additional series were constructed by shortening the vowel duration of these syllables, and the resulting series were presented to listeners in randomized sequences for identification. The results of the first series of experiments showed that shortening the duration of frication for a voiceless affricate or fricative produced a change in the percept from voiceless to voiced (i.e., from [?a] to [?a], from [fa] to [va], and from [sa] to [za]). Whereas shortening the duration of frication produced a marked shift in the percept from voiceless to voiced, shortening the duration of the vowel produced only a weak effect, and the interaction between the two shortening effects was small. The small effect of vowel duration on perception of voicing for syllable-initial affricates and fricatives was contrasted with the strong effects of vowel duration on voicing for fricatives in syllable-final position observed in prior work, and this contrast was accounted for in terms of the perceptual processing required for CVC syllables in normal listening. In the second series of experiments, the frication of [sa] was shortened to produce a series of syllables ranging from [sa] to [za] to [-a]. In selective adaptation tests, repeated presentation of these three syllables resulted in shifts in the listeners’ phonetic boundaries. Results of the selective adaptation experiment were compared to results of prior work using stop consonants. Subject Classification: 70.30, 70.70.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 68 citations exist upstream.