Automatic recognition of syllable-final nasals preceded by /ε/

P. Loizou, Michael Dorman, Andreas Spanias

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

In this paper, it is shown how an automatic recognition algorithm, based on hidden Marker models (HMM), can benefit by properly utilizing findings from perceptual experiments on nasals. Perceptual studies on nasal consonants have shown that both nasal murmurs and formant transitions are important in the identification of place of articulation. Thus both acoustic segments bordering the nasal release were incorporated into this HMM-based system. A 7% improvement in alveolar recognition was obtained by explicitly modeling the vowel-nasal transition segments. Further overall improvement (6%) was realized by making the HMM recognizer 'focus' more on the vowel-nasal transition segments bordering the nasal release, and less on the nasal murmur and vowel portion of the/ε m/and/ε n/syllables. An overall average [m]-[n] recognition of 95% was obtained when testing this technique on 60 speakers outside the training set.

Original languageEnglish (US)
Pages (from-to)1925-1928
Number of pages4
JournalJournal of the Acoustical Society of America
Volume97
Issue number3
DOIs
StatePublished - 1995

Fingerprint

syllables
vowels
markers
education
acoustics
Nasal

ASJC Scopus subject areas

  • Acoustics and Ultrasonics

Cite this

Automatic recognition of syllable-final nasals preceded by /ε/. / Loizou, P.; Dorman, Michael; Spanias, Andreas.

In: Journal of the Acoustical Society of America, Vol. 97, No. 3, 1995, p. 1925-1928.

Research output: Contribution to journalArticle

@article{44755503b6f148e0a1a3493cecaa0eb2,
title = "Automatic recognition of syllable-final nasals preceded by /ε/",
abstract = "In this paper, it is shown how an automatic recognition algorithm, based on hidden Marker models (HMM), can benefit by properly utilizing findings from perceptual experiments on nasals. Perceptual studies on nasal consonants have shown that both nasal murmurs and formant transitions are important in the identification of place of articulation. Thus both acoustic segments bordering the nasal release were incorporated into this HMM-based system. A 7{\%} improvement in alveolar recognition was obtained by explicitly modeling the vowel-nasal transition segments. Further overall improvement (6{\%}) was realized by making the HMM recognizer 'focus' more on the vowel-nasal transition segments bordering the nasal release, and less on the nasal murmur and vowel portion of the/ε m/and/ε n/syllables. An overall average [m]-[n] recognition of 95{\%} was obtained when testing this technique on 60 speakers outside the training set.",
author = "P. Loizou and Michael Dorman and Andreas Spanias",
year = "1995",
doi = "10.1121/1.412065",
language = "English (US)",
volume = "97",
pages = "1925--1928",
journal = "Journal of the Acoustical Society of America",
issn = "0001-4966",
publisher = "Acoustical Society of America",
number = "3",

}

TY - JOUR

T1 - Automatic recognition of syllable-final nasals preceded by /ε/

AU - Loizou, P.

AU - Dorman, Michael

AU - Spanias, Andreas

PY - 1995

Y1 - 1995

N2 - In this paper, it is shown how an automatic recognition algorithm, based on hidden Marker models (HMM), can benefit by properly utilizing findings from perceptual experiments on nasals. Perceptual studies on nasal consonants have shown that both nasal murmurs and formant transitions are important in the identification of place of articulation. Thus both acoustic segments bordering the nasal release were incorporated into this HMM-based system. A 7% improvement in alveolar recognition was obtained by explicitly modeling the vowel-nasal transition segments. Further overall improvement (6%) was realized by making the HMM recognizer 'focus' more on the vowel-nasal transition segments bordering the nasal release, and less on the nasal murmur and vowel portion of the/ε m/and/ε n/syllables. An overall average [m]-[n] recognition of 95% was obtained when testing this technique on 60 speakers outside the training set.

AB - In this paper, it is shown how an automatic recognition algorithm, based on hidden Marker models (HMM), can benefit by properly utilizing findings from perceptual experiments on nasals. Perceptual studies on nasal consonants have shown that both nasal murmurs and formant transitions are important in the identification of place of articulation. Thus both acoustic segments bordering the nasal release were incorporated into this HMM-based system. A 7% improvement in alveolar recognition was obtained by explicitly modeling the vowel-nasal transition segments. Further overall improvement (6%) was realized by making the HMM recognizer 'focus' more on the vowel-nasal transition segments bordering the nasal release, and less on the nasal murmur and vowel portion of the/ε m/and/ε n/syllables. An overall average [m]-[n] recognition of 95% was obtained when testing this technique on 60 speakers outside the training set.

UR - http://www.scopus.com/inward/record.url?scp=0028936631&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0028936631&partnerID=8YFLogxK

U2 - 10.1121/1.412065

DO - 10.1121/1.412065

M3 - Article

C2 - 7699174

AN - SCOPUS:0028936631

VL - 97

SP - 1925

EP - 1928

JO - Journal of the Acoustical Society of America

JF - Journal of the Acoustical Society of America

SN - 0001-4966

IS - 3

ER -