Frequency-domain delexicalization using surrogate vowels

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

We propose a delexicalization algorithm that renders the lexical content of an utterance unintelligible, while preserving important acoustic prosodic cues, as well as naturalness and speaker identity. This is achieved by replacing voiced regions by spectral slices from a surrogate vowel, and by averaging the magnitude spectrum during unvoiced regions. Perceptual tests were carried out comparing sentences that were either unprocessed or delexicalized, using a baseline or the proposed method. An intelligibility test resulted in a keyword recall rate of 92% for the unprocessed sentences, and near complete unintelligibility for both delexicalization methods. Affect recognition was at 65% for unprocessed sentences, and 46% and 49% for the baseline and the proposed method, respectively. Preference tests showed that the proposed method preserved drastically more speaker identity, and sounded more natural than the baseline.

Original languageEnglish (US)
Title of host publicationProceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010
Pages474-477
Number of pages4
StatePublished - 2010
Event11th Annual Conference of the International Speech Communication Association: Spoken Language Processing for All, INTERSPEECH 2010 - Makuhari, Chiba, Japan
Duration: Sep 26 2010Sep 30 2010

Other

Other11th Annual Conference of the International Speech Communication Association: Spoken Language Processing for All, INTERSPEECH 2010
CountryJapan
CityMakuhari, Chiba
Period9/26/109/30/10

Fingerprint

Acoustics
Cues
Recognition (Psychology)
Unvoiced
Key Words
Naturalness
Render
Intelligibility
Spectrality
Unintelligibility
Utterance

Keywords

  • Affect
  • Delexicalization
  • Intelligibility

ASJC Scopus subject areas

  • Language and Linguistics
  • Speech and Hearing

Cite this

Kain, A., & Van Santen, J. (2010). Frequency-domain delexicalization using surrogate vowels. In Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010 (pp. 474-477)

Frequency-domain delexicalization using surrogate vowels. / Kain, Alexander; Van Santen, Jan.

Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. 2010. p. 474-477.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Kain, A & Van Santen, J 2010, Frequency-domain delexicalization using surrogate vowels. in Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. pp. 474-477, 11th Annual Conference of the International Speech Communication Association: Spoken Language Processing for All, INTERSPEECH 2010, Makuhari, Chiba, Japan, 9/26/10.
Kain A, Van Santen J. Frequency-domain delexicalization using surrogate vowels. In Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. 2010. p. 474-477
Kain, Alexander ; Van Santen, Jan. / Frequency-domain delexicalization using surrogate vowels. Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. 2010. pp. 474-477
@inproceedings{2624cae27f0848e5858917d1fdf8d2a7,
title = "Frequency-domain delexicalization using surrogate vowels",
abstract = "We propose a delexicalization algorithm that renders the lexical content of an utterance unintelligible, while preserving important acoustic prosodic cues, as well as naturalness and speaker identity. This is achieved by replacing voiced regions by spectral slices from a surrogate vowel, and by averaging the magnitude spectrum during unvoiced regions. Perceptual tests were carried out comparing sentences that were either unprocessed or delexicalized, using a baseline or the proposed method. An intelligibility test resulted in a keyword recall rate of 92{\%} for the unprocessed sentences, and near complete unintelligibility for both delexicalization methods. Affect recognition was at 65{\%} for unprocessed sentences, and 46{\%} and 49{\%} for the baseline and the proposed method, respectively. Preference tests showed that the proposed method preserved drastically more speaker identity, and sounded more natural than the baseline.",
keywords = "Affect, Delexicalization, Intelligibility",
author = "Alexander Kain and {Van Santen}, Jan",
year = "2010",
language = "English (US)",
pages = "474--477",
booktitle = "Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010",

}

TY - GEN

T1 - Frequency-domain delexicalization using surrogate vowels

AU - Kain, Alexander

AU - Van Santen, Jan

PY - 2010

Y1 - 2010

N2 - We propose a delexicalization algorithm that renders the lexical content of an utterance unintelligible, while preserving important acoustic prosodic cues, as well as naturalness and speaker identity. This is achieved by replacing voiced regions by spectral slices from a surrogate vowel, and by averaging the magnitude spectrum during unvoiced regions. Perceptual tests were carried out comparing sentences that were either unprocessed or delexicalized, using a baseline or the proposed method. An intelligibility test resulted in a keyword recall rate of 92% for the unprocessed sentences, and near complete unintelligibility for both delexicalization methods. Affect recognition was at 65% for unprocessed sentences, and 46% and 49% for the baseline and the proposed method, respectively. Preference tests showed that the proposed method preserved drastically more speaker identity, and sounded more natural than the baseline.

AB - We propose a delexicalization algorithm that renders the lexical content of an utterance unintelligible, while preserving important acoustic prosodic cues, as well as naturalness and speaker identity. This is achieved by replacing voiced regions by spectral slices from a surrogate vowel, and by averaging the magnitude spectrum during unvoiced regions. Perceptual tests were carried out comparing sentences that were either unprocessed or delexicalized, using a baseline or the proposed method. An intelligibility test resulted in a keyword recall rate of 92% for the unprocessed sentences, and near complete unintelligibility for both delexicalization methods. Affect recognition was at 65% for unprocessed sentences, and 46% and 49% for the baseline and the proposed method, respectively. Preference tests showed that the proposed method preserved drastically more speaker identity, and sounded more natural than the baseline.

KW - Affect

KW - Delexicalization

KW - Intelligibility

UR - http://www.scopus.com/inward/record.url?scp=79959825126&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79959825126&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:79959825126

SP - 474

EP - 477

BT - Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010

ER -