1981The Journal of the Acoustical Society of AmericaRequires access

Second formants in fricatives: Acoustic consequences of fricative-vowel coarticulation

Sigfrid D. Soli

Open publisher page 141 citations

Abstract

Spectral analyses of the sibilant fricatives, [s, z, ℱ, z], as spoken in isolation and in initial position before the vowels [a, i, u], were performed for a corpus of 80 utterances. The mean LPC spectra for five utterances of each fricative produced in each vowel context revealed reliable anticipatory vowel coarticulation effects present at least 30–60 ms before vowel onset in the form of spectral peaks affiliated with the second formant of the vowel. Peak frequencies were between 1.5 and 2 kHz and were approximately 100–300 Hz higher before the front vowel [i] than before the back vowels [a, u]. The presence of these peaks in the fricative spectra indicate that during the latter part of the fricative the constriction begins to open in anticipation of the vowel, and, as a consequence, second formant resonances are excited. The acoustic characteristics of these fricative formants varied as a function of vowel context due to differences in the anticipatory coarticulation of the vowel. Fricative place of constriction was assimilated to the high vowels [i, u], resulting in different fricative formant frequencies, while overlap of fricative and vowel articulatory gestures produced additional acoustic differences. These effects were least in the [a] context because the opposing gestures required to produce the constricted fricative and the relatively open low vowel are presumably performed in sequence with minimal assimilation or overlap. The present acoustic evidence and related articulatory considerations are congruent with the results from a companion perceptual study [G. H. Yeni–Komshian and S. D. Soli, J. Acoust. Soc. Am. 70, 966–975 (1981)] showing that the high vowels [i, u] but not the low vowel [a], can be identified in fricatives excised from fricative–vowel syllables. These objective links between articulatory, acoustic, and perceptual aspects of coarticulation are discussed.

About this research paper

What this paper is about

Spectral analyses of the sibilant fricatives, [s, z, ℱ, z], as spoken in isolation and in initial position before the vowels [a, i, u], were performed for a corpus of 80 utterances. The mean LPC spectra for five utterances of each fricative produced in each vowel context revealed reliable anticipatory vowel coarticulation effects present at least 30–60 ms before vowel onset in the form of spectral peaks affiliated with the second formant of the vowel. Peak frequencies were between 1.5 and 2 kHz and were approximately 100–300 Hz higher before the front vowel [i] than before the back vowels [a, u]. The presence of these peaks in the fricative spectra indicate that during the latter part of the fricative the constriction begins to open in anticipation of the vowel, and, as a consequence, second formant resonances are excited. The acoustic characteristics of these fricative formants varied as a function of vowel context due to differences in the anticipatory coarticulation of the vowel. Fricative place of constriction was assimilated to the high vowels [i, u], resulting in different fricative formant frequencies, while overlap of fricative and vowel articulatory gestures produced additional acoustic differences. These effects were least in the [a] context because the opposing gestures required to produce the constricted fricative and the relatively open low vowel are presumably performed in sequence with minimal assimilation or overlap. The present acoustic evidence and related articulatory considerations are congruent with the results from a companion perceptual study [G. H. Yeni–Komshian and S. D. Soli, J. Acoust. Soc. Am. 70, 966–975 (1981)] showing that the high vowels [i, u] but not the low vowel [a], can be identified in fricatives excised from fricative–vowel syllables. These objective links between articulatory, acoustic, and perceptual aspects of coarticulation are discussed.

Why it matters

OpenAlex reports 141 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Spectral analyses of the sibilant fricatives, [s, z, ℱ, z], as spoken in isolation and in initial position before the vowels [a, i, u], were performed for a corpus of 80 utterances. The mean LPC spectra for five utterances of each fricative produced in each vowel context revealed reliable anticipatory vowel coarticulation effects present at least 30–60 ms before vowel onset in the form of spectral peaks affiliated with the second formant of the vowel. Peak frequencies were between 1.5 and 2 kHz and were approximately 100–300 Hz higher before the front vowel [i] than before the back vowels [a, u]. The presence of these peaks in the fricative spectra indicate that during the latter part of the fricative the constriction begins to open in anticipation of the vowel, and, as a consequence, second formant resonances are excited. The acoustic characteristics of these fricative formants varied as a function of vowel context due to differences in the anticipatory coarticulation of the vowel. Fricative place of constriction was assimilated to the high vowels [i, u], resulting in different fricative formant frequencies, while overlap of fricative and vowel articulatory gestures produced additional acoustic differences. These effects were least in the [a] context because the opposing gestures required to produce the constricted fricative and the relatively open low vowel are presumably performed in sequence with minimal assimilation or overlap. The present acoustic evidence and related articulatory considerations are congruent with the results from a companion perceptual study [G. H. Yeni–Komshian and S. D. Soli, J. Acoust. Soc. Am. 70, 966–975 (1981)] showing that the high vowels [i, u] but not the low vowel [a], can be identified in fricatives excised from fricative–vowel syllables. These objective links between articulatory, acoustic, and perceptual aspects of coarticulation are discussed.

Key concepts: Formant, Coarticulation, Vowel, Acoustics, Context (archaeology), Mathematics, Mid vowel, Speech recognition

Related papers

Back to paper searchBrowse research topicsOriginal source
Second formants in fricatives: Acoustic consequences of fricative-vowel coarticulation — Research Paper | ScholarLens