Cepstral analysis based on the Glimpse proportion measure for improving the intelligibility of HMM-based synthetic speech in noise

C. Valentini-Botinhao; R. Maia; J. Yamagishi; S. King; H. Zen

doi:10.1109/ICASSP.2012.6288794

Cepstral analysis based on the Glimpse proportion measure for improving the intelligibility of HMM-based synthetic speech in noise

C. Valentini-Botinhao, R. Maia, J. Yamagishi, S. King, H. Zen

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

Abstract

In this paper we introduce a new cepstral coefficient extraction method based on an intelligibility measure for speech in noise, the Glimpse Proportion measure. This new method aims to increase the intelligibility of speech in noise by modifying the clean speech, and has applications in scenarios such as public announcement and car navigation systems. We first explain how the Glimpse Proportion measure operates and further show how we approximated it to integrate it into an existing spectral envelope parameter extraction method commonly used in the HMM-based speech synthesis framework. We then demonstrate how this new method changes the modelled spectrum according to the characteristics of the noise and show results for a listening test with vocoded and HMM-based synthetic speech. The test indicates that the proposed method can significantly improve intelligibility of synthetic speech in speech shaped noise.

Original language	English
Title of host publication	2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP)
Subtitle of host publication	Kyoto, Japan
Place of Publication	NEW YORK
Publisher	Institute of Electrical and Electronics Engineers
Pages	3997-4000
Number of pages	4
ISBN (Electronic)	978-1-4673-0044-5
ISBN (Print)	978-1-4673-0045-2
DOIs	https://doi.org/10.1109/ICASSP.2012.6288794
Publication status	Published - 2012
Event	IEEE International Conference on Acoustics, Speech and Signal Processing - Kyoto Duration: 25 Mar 2012 → 30 Mar 2012

Publication series

Name	IEEE International Conference on Acoustics Speech and Signal Processing
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
ISSN (Print)	1520-6149

Conference

Conference	IEEE International Conference on Acoustics, Speech and Signal Processing
City	Kyoto
Period	25/03/12 → 30/03/12

Keywords / Materials (for Non-textual outputs)

approximation theory
hidden Markov models
speech coding
speech synthesis
HMM-based speech synthesis framework
HMM-based synthetic speech intelligibility
car navigation systems
cepstral coefficient extraction method
clean speech modification
glimpse proportion measure
speech shaped noise
vocoded synthetic speech
Accuracy
Approximation methods
Cepstral analysis
Hidden Markov models
Noise
Noise measurement
Speech
HMM-based speech synthesis
Lombard speech
cepstral coefficient extraction
objective measure for speech intelligibility

Access to Document

10.1109/ICASSP.2012.6288794

Cepstral analysis based on the Glimpse Proportion measure for improving the intelligibility of HMM-based synthetic speech in noise
alentini-Botinhao, C., Maia, R., Yamagishi, J., King, S., & Zen, H. (2012). Cepstral analysis based on the Glimpse proportion measure for improving the intelligibility of HMM-based synthetic speech in noise. In 2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP). (pp. 3997-4000). (IEEE International Conference on Acoustics Speech and Signal Processing). NEW YORK: IEEE. doi: 10.1109/ICASSP.2012.6288794
Accepted author manuscript, 253 KB

Deep architectures for statistical speech synthesis
Yamagishi, J. (Principal Investigator)
UK industry, commerce and public corporations
4/09/12 → 3/03/16
Project: Research
LISTA: LISTA- The Listening Talker (RTGS)
King, S. (Principal Investigator), Mayo, C. (Co-investigator) & Renals, S. (Co-investigator)
EU government bodies
1/05/10 → 30/04/13
Project: Research

EACL 2014 keynote: Speech synthesis needs YOU!
King, S. (Speaker)
29 Apr 2014
Activity: Academic talk or presentation types › Invited talk

File

Cite this

Valentini-Botinhao, C., Maia, R., Yamagishi, J., King, S., & Zen, H. (2012). Cepstral analysis based on the Glimpse proportion measure for improving the intelligibility of HMM-based synthetic speech in noise. In 2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP): Kyoto, Japan (pp. 3997-4000). (IEEE International Conference on Acoustics Speech and Signal Processing). Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ICASSP.2012.6288794

Valentini-Botinhao, C. ; Maia, R. ; Yamagishi, J. et al. / Cepstral analysis based on the Glimpse proportion measure for improving the intelligibility of HMM-based synthetic speech in noise. 2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP): Kyoto, Japan. NEW YORK : Institute of Electrical and Electronics Engineers, 2012. pp. 3997-4000 (IEEE International Conference on Acoustics Speech and Signal Processing).

@inproceedings{84010373835a4ca8893a95e12b83cb4f,

title = "Cepstral analysis based on the Glimpse proportion measure for improving the intelligibility of HMM-based synthetic speech in noise",

abstract = "In this paper we introduce a new cepstral coefficient extraction method based on an intelligibility measure for speech in noise, the Glimpse Proportion measure. This new method aims to increase the intelligibility of speech in noise by modifying the clean speech, and has applications in scenarios such as public announcement and car navigation systems. We first explain how the Glimpse Proportion measure operates and further show how we approximated it to integrate it into an existing spectral envelope parameter extraction method commonly used in the HMM-based speech synthesis framework. We then demonstrate how this new method changes the modelled spectrum according to the characteristics of the noise and show results for a listening test with vocoded and HMM-based synthetic speech. The test indicates that the proposed method can significantly improve intelligibility of synthetic speech in speech shaped noise.",

keywords = "approximation theory, hidden Markov models, speech coding, speech synthesis, HMM-based speech synthesis framework, HMM-based synthetic speech intelligibility, car navigation systems, cepstral coefficient extraction method, clean speech modification, glimpse proportion measure, speech shaped noise, vocoded synthetic speech, Accuracy, Approximation methods, Cepstral analysis, Hidden Markov models, Noise, Noise measurement, Speech, HMM-based speech synthesis, Lombard speech, cepstral coefficient extraction, objective measure for speech intelligibility",

author = "C. Valentini-Botinhao and R. Maia and J. Yamagishi and S. King and H. Zen",

year = "2012",

doi = "10.1109/ICASSP.2012.6288794",

language = "English",

isbn = "978-1-4673-0045-2",

series = "IEEE International Conference on Acoustics Speech and Signal Processing",

publisher = "Institute of Electrical and Electronics Engineers",

pages = "3997--4000",

booktitle = "2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP)",

address = "United States",

note = "IEEE International Conference on Acoustics, Speech and Signal Processing ; Conference date: 25-03-2012 Through 30-03-2012",

}

Valentini-Botinhao, C, Maia, R, Yamagishi, J, King, S & Zen, H 2012, Cepstral analysis based on the Glimpse proportion measure for improving the intelligibility of HMM-based synthetic speech in noise. in 2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP): Kyoto, Japan. IEEE International Conference on Acoustics Speech and Signal Processing, Institute of Electrical and Electronics Engineers, NEW YORK, pp. 3997-4000, IEEE International Conference on Acoustics, Speech and Signal Processing, Kyoto, 25/03/12. https://doi.org/10.1109/ICASSP.2012.6288794

Cepstral analysis based on the Glimpse proportion measure for improving the intelligibility of HMM-based synthetic speech in noise. / Valentini-Botinhao, C.; Maia, R.; Yamagishi, J. et al.
2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP): Kyoto, Japan. NEW YORK: Institute of Electrical and Electronics Engineers, 2012. p. 3997-4000 (IEEE International Conference on Acoustics Speech and Signal Processing).

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

TY - GEN

T1 - Cepstral analysis based on the Glimpse proportion measure for improving the intelligibility of HMM-based synthetic speech in noise

AU - Valentini-Botinhao, C.

AU - Maia, R.

AU - Yamagishi, J.

AU - King, S.

AU - Zen, H.

PY - 2012

Y1 - 2012

N2 - In this paper we introduce a new cepstral coefficient extraction method based on an intelligibility measure for speech in noise, the Glimpse Proportion measure. This new method aims to increase the intelligibility of speech in noise by modifying the clean speech, and has applications in scenarios such as public announcement and car navigation systems. We first explain how the Glimpse Proportion measure operates and further show how we approximated it to integrate it into an existing spectral envelope parameter extraction method commonly used in the HMM-based speech synthesis framework. We then demonstrate how this new method changes the modelled spectrum according to the characteristics of the noise and show results for a listening test with vocoded and HMM-based synthetic speech. The test indicates that the proposed method can significantly improve intelligibility of synthetic speech in speech shaped noise.

AB - In this paper we introduce a new cepstral coefficient extraction method based on an intelligibility measure for speech in noise, the Glimpse Proportion measure. This new method aims to increase the intelligibility of speech in noise by modifying the clean speech, and has applications in scenarios such as public announcement and car navigation systems. We first explain how the Glimpse Proportion measure operates and further show how we approximated it to integrate it into an existing spectral envelope parameter extraction method commonly used in the HMM-based speech synthesis framework. We then demonstrate how this new method changes the modelled spectrum according to the characteristics of the noise and show results for a listening test with vocoded and HMM-based synthetic speech. The test indicates that the proposed method can significantly improve intelligibility of synthetic speech in speech shaped noise.

KW - approximation theory

KW - hidden Markov models

KW - speech coding

KW - speech synthesis

KW - HMM-based speech synthesis framework

KW - HMM-based synthetic speech intelligibility

KW - car navigation systems

KW - cepstral coefficient extraction method

KW - clean speech modification

KW - glimpse proportion measure

KW - speech shaped noise

KW - vocoded synthetic speech

KW - Accuracy

KW - Approximation methods

KW - Cepstral analysis

KW - Hidden Markov models

KW - Noise

KW - Noise measurement

KW - Speech

KW - HMM-based speech synthesis

KW - Lombard speech

KW - cepstral coefficient extraction

KW - objective measure for speech intelligibility

U2 - 10.1109/ICASSP.2012.6288794

DO - 10.1109/ICASSP.2012.6288794

M3 - Conference contribution

SN - 978-1-4673-0045-2

T3 - IEEE International Conference on Acoustics Speech and Signal Processing

SP - 3997

EP - 4000

BT - 2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP)

PB - Institute of Electrical and Electronics Engineers

CY - NEW YORK

T2 - IEEE International Conference on Acoustics, Speech and Signal Processing

Y2 - 25 March 2012 through 30 March 2012

ER -

Valentini-Botinhao C, Maia R, Yamagishi J, King S, Zen H. Cepstral analysis based on the Glimpse proportion measure for improving the intelligibility of HMM-based synthetic speech in noise. In 2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP): Kyoto, Japan. NEW YORK: Institute of Electrical and Electronics Engineers. 2012. p. 3997-4000. (IEEE International Conference on Acoustics Speech and Signal Processing). doi: 10.1109/ICASSP.2012.6288794

Cepstral analysis based on the Glimpse proportion measure for improving the intelligibility of HMM-based synthetic speech in noise

Abstract

Publication series

Conference

Keywords / Materials (for Non-textual outputs)

Access to Document

Fingerprint

Projects

Deep architectures for statistical speech synthesis

LISTA: LISTA- The Listening Talker (RTGS)

Activities

EACL 2014 keynote: Speech synthesis needs YOU!

Cite this