Vocal source features for bilingual speaker identification

Jianglin Wang; Michael T. Johnson

doi:10.1109/ChinaSIP.2013.6625321

Vocal source features for bilingual speaker identification

Jianglin Wang, Michael T. Johnson

Electrical and Computer Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

6 Scopus citations

Abstract

This paper introduces the use of two new features for speaker identification, Residual Phase Cepstrum Coefficients (RPCC) and Glottal Flow Cepstrum Coefficients (GLFCC), to capture speaker-specific characteristics from their vocal excitation patterns. Results on a cross-lingual speaker identification task taken from the NIST 2004 SRE demonstrate that these RPCC and GLFCC features are significantly more accurate than traditional mel-frequency cepstral coefficients (MFCC). In particular, these two new features give better results with smaller amounts of training data, due to lower model complexity.

Original language	English
Title of host publication	2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings
Pages	170-173
Number of pages	4
DOIs	https://doi.org/10.1109/ChinaSIP.2013.6625321
State	Published - 2013
Event	2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Beijing, China Duration: Jul 6 2013 → Jul 10 2013

Publication series

Name	2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings

Conference

Conference	2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013
Country/Territory	China
City	Beijing
Period	7/6/13 → 7/10/13

Keywords

Glottal source excitation
IAIF and GMM
Speaker identification

ASJC Scopus subject areas

Information Systems
Signal Processing

Access to Document

10.1109/ChinaSIP.2013.6625321

Cite this

Wang, J., & Johnson, M. T. (2013). Vocal source features for bilingual speaker identification. In 2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings (pp. 170-173). Article 6625321 (2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings). https://doi.org/10.1109/ChinaSIP.2013.6625321

@inproceedings{fea005d8fa554f48b441954ec5b7f19a,

title = "Vocal source features for bilingual speaker identification",

abstract = "This paper introduces the use of two new features for speaker identification, Residual Phase Cepstrum Coefficients (RPCC) and Glottal Flow Cepstrum Coefficients (GLFCC), to capture speaker-specific characteristics from their vocal excitation patterns. Results on a cross-lingual speaker identification task taken from the NIST 2004 SRE demonstrate that these RPCC and GLFCC features are significantly more accurate than traditional mel-frequency cepstral coefficients (MFCC). In particular, these two new features give better results with smaller amounts of training data, due to lower model complexity.",

keywords = "Glottal source excitation, IAIF and GMM, Speaker identification",

author = "Jianglin Wang and Johnson, {Michael T.}",

year = "2013",

doi = "10.1109/ChinaSIP.2013.6625321",

language = "English",

isbn = "9781479910434",

series = "2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings",

pages = "170--173",

booktitle = "2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings",

note = "2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 ; Conference date: 06-07-2013 Through 10-07-2013",

}

Wang, J & Johnson, MT 2013, Vocal source features for bilingual speaker identification. in 2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings., 6625321, 2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings, pp. 170-173, 2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013, Beijing, China, 7/6/13. https://doi.org/10.1109/ChinaSIP.2013.6625321

Vocal source features for bilingual speaker identification. / Wang, Jianglin; Johnson, Michael T.
2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings. 2013. p. 170-173 6625321 (2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Vocal source features for bilingual speaker identification

AU - Wang, Jianglin

AU - Johnson, Michael T.

PY - 2013

Y1 - 2013

N2 - This paper introduces the use of two new features for speaker identification, Residual Phase Cepstrum Coefficients (RPCC) and Glottal Flow Cepstrum Coefficients (GLFCC), to capture speaker-specific characteristics from their vocal excitation patterns. Results on a cross-lingual speaker identification task taken from the NIST 2004 SRE demonstrate that these RPCC and GLFCC features are significantly more accurate than traditional mel-frequency cepstral coefficients (MFCC). In particular, these two new features give better results with smaller amounts of training data, due to lower model complexity.

AB - This paper introduces the use of two new features for speaker identification, Residual Phase Cepstrum Coefficients (RPCC) and Glottal Flow Cepstrum Coefficients (GLFCC), to capture speaker-specific characteristics from their vocal excitation patterns. Results on a cross-lingual speaker identification task taken from the NIST 2004 SRE demonstrate that these RPCC and GLFCC features are significantly more accurate than traditional mel-frequency cepstral coefficients (MFCC). In particular, these two new features give better results with smaller amounts of training data, due to lower model complexity.

KW - Glottal source excitation

KW - IAIF and GMM

KW - Speaker identification

UR - http://www.scopus.com/inward/record.url?scp=84889586711&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84889586711&partnerID=8YFLogxK

U2 - 10.1109/ChinaSIP.2013.6625321

DO - 10.1109/ChinaSIP.2013.6625321

M3 - Conference contribution

AN - SCOPUS:84889586711

SN - 9781479910434

T3 - 2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings

SP - 170

EP - 173

BT - 2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings

T2 - 2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013

Y2 - 6 July 2013 through 10 July 2013

ER -

Wang J, Johnson MT. Vocal source features for bilingual speaker identification. In 2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings. 2013. p. 170-173. 6625321. (2013 IEEE China Summit and International Conference on Signal and Information Processing, ChinaSIP 2013 - Proceedings). doi: 10.1109/ChinaSIP.2013.6625321

Vocal source features for bilingual speaker identification

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this