Estimation of the vocal tract spectrum from articulatory movements using phoneme-dependent neural networks

Kohei Wakamiya, Tokihiko Kaburagi, Takuya Tsuji, Jiji Kim

Research output: Contribution to conferencePaper

Abstract

This paper presents an estimation method of the vocal tract spectrum from articulatory movements. The method is based on the interpolation of spectra obtained by phoneme-dependent neural networks. Given the phonemic context and articulation timing corresponding to each phoneme, the proposed method first transforms articulator positions to phoneme-dependent spectra. Then the vocal tract spectrum is estimated by the interpolation of transformed spectra. This interpolation is based on the distance among the input articulator position and that of the preceding and succeeding phonemes. Also, a training procedure of the neural networks is presented while taking the spectral interpolation into account. Articulatory and acoustic data pairs collected by a simultaneous recording of articulator positions and speech were used as the training and test data. Finally, we showed an estimation result using the proposed method.

Original languageEnglish
Pages517-520
Number of pages4
Publication statusPublished - Jan 1 2004
Event8th International Conference on Spoken Language Processing, ICSLP 2004 - Jeju, Jeju Island, Korea, Republic of
Duration: Oct 4 2004Oct 8 2004

Other

Other8th International Conference on Spoken Language Processing, ICSLP 2004
CountryKorea, Republic of
CityJeju, Jeju Island
Period10/4/0410/8/04

Fingerprint

neural network
acoustics
recording
Neural Networks
Vocal Tract
Phoneme
Interpolation
Articulator

All Science Journal Classification (ASJC) codes

  • Language and Linguistics
  • Linguistics and Language

Cite this

Wakamiya, K., Kaburagi, T., Tsuji, T., & Kim, J. (2004). Estimation of the vocal tract spectrum from articulatory movements using phoneme-dependent neural networks. 517-520. Paper presented at 8th International Conference on Spoken Language Processing, ICSLP 2004, Jeju, Jeju Island, Korea, Republic of.

Estimation of the vocal tract spectrum from articulatory movements using phoneme-dependent neural networks. / Wakamiya, Kohei; Kaburagi, Tokihiko; Tsuji, Takuya; Kim, Jiji.

2004. 517-520 Paper presented at 8th International Conference on Spoken Language Processing, ICSLP 2004, Jeju, Jeju Island, Korea, Republic of.

Research output: Contribution to conferencePaper

Wakamiya, K, Kaburagi, T, Tsuji, T & Kim, J 2004, 'Estimation of the vocal tract spectrum from articulatory movements using phoneme-dependent neural networks', Paper presented at 8th International Conference on Spoken Language Processing, ICSLP 2004, Jeju, Jeju Island, Korea, Republic of, 10/4/04 - 10/8/04 pp. 517-520.
Wakamiya K, Kaburagi T, Tsuji T, Kim J. Estimation of the vocal tract spectrum from articulatory movements using phoneme-dependent neural networks. 2004. Paper presented at 8th International Conference on Spoken Language Processing, ICSLP 2004, Jeju, Jeju Island, Korea, Republic of.
Wakamiya, Kohei ; Kaburagi, Tokihiko ; Tsuji, Takuya ; Kim, Jiji. / Estimation of the vocal tract spectrum from articulatory movements using phoneme-dependent neural networks. Paper presented at 8th International Conference on Spoken Language Processing, ICSLP 2004, Jeju, Jeju Island, Korea, Republic of.4 p.
@conference{c8d08484282b44cca61b867762b0036c,
title = "Estimation of the vocal tract spectrum from articulatory movements using phoneme-dependent neural networks",
abstract = "This paper presents an estimation method of the vocal tract spectrum from articulatory movements. The method is based on the interpolation of spectra obtained by phoneme-dependent neural networks. Given the phonemic context and articulation timing corresponding to each phoneme, the proposed method first transforms articulator positions to phoneme-dependent spectra. Then the vocal tract spectrum is estimated by the interpolation of transformed spectra. This interpolation is based on the distance among the input articulator position and that of the preceding and succeeding phonemes. Also, a training procedure of the neural networks is presented while taking the spectral interpolation into account. Articulatory and acoustic data pairs collected by a simultaneous recording of articulator positions and speech were used as the training and test data. Finally, we showed an estimation result using the proposed method.",
author = "Kohei Wakamiya and Tokihiko Kaburagi and Takuya Tsuji and Jiji Kim",
year = "2004",
month = "1",
day = "1",
language = "English",
pages = "517--520",
note = "8th International Conference on Spoken Language Processing, ICSLP 2004 ; Conference date: 04-10-2004 Through 08-10-2004",

}

TY - CONF

T1 - Estimation of the vocal tract spectrum from articulatory movements using phoneme-dependent neural networks

AU - Wakamiya, Kohei

AU - Kaburagi, Tokihiko

AU - Tsuji, Takuya

AU - Kim, Jiji

PY - 2004/1/1

Y1 - 2004/1/1

N2 - This paper presents an estimation method of the vocal tract spectrum from articulatory movements. The method is based on the interpolation of spectra obtained by phoneme-dependent neural networks. Given the phonemic context and articulation timing corresponding to each phoneme, the proposed method first transforms articulator positions to phoneme-dependent spectra. Then the vocal tract spectrum is estimated by the interpolation of transformed spectra. This interpolation is based on the distance among the input articulator position and that of the preceding and succeeding phonemes. Also, a training procedure of the neural networks is presented while taking the spectral interpolation into account. Articulatory and acoustic data pairs collected by a simultaneous recording of articulator positions and speech were used as the training and test data. Finally, we showed an estimation result using the proposed method.

AB - This paper presents an estimation method of the vocal tract spectrum from articulatory movements. The method is based on the interpolation of spectra obtained by phoneme-dependent neural networks. Given the phonemic context and articulation timing corresponding to each phoneme, the proposed method first transforms articulator positions to phoneme-dependent spectra. Then the vocal tract spectrum is estimated by the interpolation of transformed spectra. This interpolation is based on the distance among the input articulator position and that of the preceding and succeeding phonemes. Also, a training procedure of the neural networks is presented while taking the spectral interpolation into account. Articulatory and acoustic data pairs collected by a simultaneous recording of articulator positions and speech were used as the training and test data. Finally, we showed an estimation result using the proposed method.

UR - http://www.scopus.com/inward/record.url?scp=85009083774&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85009083774&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:85009083774

SP - 517

EP - 520

ER -