Vowel imitation using vocal tract model and recurrent neural network

Hisashi Kanda; Tetsuya Ogata; Kazunori Komatani; Hiroshi G. Okuno

doi:10.1007/978-3-540-69162-4_24

Vowel imitation using vocal tract model and recurrent neural network

Hisashi Kanda^*, Tetsuya Ogata, Kazunori Komatani, Hiroshi G. Okuno

^*Corresponding author for this work

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

1 Citation (Scopus)

Abstract

A vocal imitation system was developed using a computational model that supports the motor theory of speech perception. A critical problem in vocal imitation is how to generate speech sounds produced by adults, whose vocal tracts have physical properties (i.e., articulatory motions) differing from those of infants' vocal tracts. To solve this problem, a model based on the motor theory of speech perception, was constructed. Applying this model enables the vocal imitation system to estimate articulatory motions for unexperienced speech sounds that have not actually been generated by the system. The system was implemented by using Recurrent Neural Network with Parametric Bias (RNNPB) and a physical vocal tract model, called Maeda model. Experimental results demonstrated that the system was sufficiently robust with respect to individual differences in speech sounds and could imitate unexperienced vowel sounds.

Original language	English
Title of host publication	Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers
Pages	222-232
Number of pages	11
Edition	PART 2
DOIs	https://doi.org/10.1007/978-3-540-69162-4_24
Publication status	Published - 2008 Oct 23
Externally published	Yes
Event	14th International Conference on Neural Information Processing, ICONIP 2007 - Kitakyushu, Japan Duration: 2007 Nov 13 → 2007 Nov 16

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 2
Volume	4985 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	14th International Conference on Neural Information Processing, ICONIP 2007
Country/Territory	Japan
City	Kitakyushu
Period	07/11/13 → 07/11/16

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-540-69162-4_24

Cite this

Kanda, H., Ogata, T., Komatani, K., & Okuno, H. G. (2008). Vowel imitation using vocal tract model and recurrent neural network. In Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers (PART 2 ed., pp. 222-232). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4985 LNCS, No. PART 2). https://doi.org/10.1007/978-3-540-69162-4_24

Vowel imitation using vocal tract model and recurrent neural network. / Kanda, Hisashi; Ogata, Tetsuya; Komatani, Kazunori et al.
Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers. PART 2. ed. 2008. p. 222-232 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4985 LNCS, No. PART 2).

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

Kanda, H, Ogata, T, Komatani, K & Okuno, HG 2008, Vowel imitation using vocal tract model and recurrent neural network. in Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers. PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 4985 LNCS, pp. 222-232, 14th International Conference on Neural Information Processing, ICONIP 2007, Kitakyushu, Japan, 07/11/13. https://doi.org/10.1007/978-3-540-69162-4_24

Kanda H, Ogata T, Komatani K, Okuno HG. Vowel imitation using vocal tract model and recurrent neural network. In Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers. PART 2 ed. 2008. p. 222-232. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). doi: 10.1007/978-3-540-69162-4_24

Kanda, Hisashi ; Ogata, Tetsuya ; Komatani, Kazunori et al. / Vowel imitation using vocal tract model and recurrent neural network. Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers. PART 2. ed. 2008. pp. 222-232 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).

@inproceedings{0522f94525324dbda2494fe37da44aab,

title = "Vowel imitation using vocal tract model and recurrent neural network",

abstract = "A vocal imitation system was developed using a computational model that supports the motor theory of speech perception. A critical problem in vocal imitation is how to generate speech sounds produced by adults, whose vocal tracts have physical properties (i.e., articulatory motions) differing from those of infants' vocal tracts. To solve this problem, a model based on the motor theory of speech perception, was constructed. Applying this model enables the vocal imitation system to estimate articulatory motions for unexperienced speech sounds that have not actually been generated by the system. The system was implemented by using Recurrent Neural Network with Parametric Bias (RNNPB) and a physical vocal tract model, called Maeda model. Experimental results demonstrated that the system was sufficiently robust with respect to individual differences in speech sounds and could imitate unexperienced vowel sounds.",

author = "Hisashi Kanda and Tetsuya Ogata and Kazunori Komatani and Okuno, {Hiroshi G.}",

year = "2008",

month = oct,

day = "23",

doi = "10.1007/978-3-540-69162-4_24",

language = "English",

isbn = "3540691596",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

number = "PART 2",

pages = "222--232",

booktitle = "Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers",

edition = "PART 2",

note = "14th International Conference on Neural Information Processing, ICONIP 2007 ; Conference date: 13-11-2007 Through 16-11-2007",

}

TY - GEN

T1 - Vowel imitation using vocal tract model and recurrent neural network

AU - Kanda, Hisashi

AU - Ogata, Tetsuya

AU - Komatani, Kazunori

AU - Okuno, Hiroshi G.

PY - 2008/10/23

Y1 - 2008/10/23

N2 - A vocal imitation system was developed using a computational model that supports the motor theory of speech perception. A critical problem in vocal imitation is how to generate speech sounds produced by adults, whose vocal tracts have physical properties (i.e., articulatory motions) differing from those of infants' vocal tracts. To solve this problem, a model based on the motor theory of speech perception, was constructed. Applying this model enables the vocal imitation system to estimate articulatory motions for unexperienced speech sounds that have not actually been generated by the system. The system was implemented by using Recurrent Neural Network with Parametric Bias (RNNPB) and a physical vocal tract model, called Maeda model. Experimental results demonstrated that the system was sufficiently robust with respect to individual differences in speech sounds and could imitate unexperienced vowel sounds.

AB - A vocal imitation system was developed using a computational model that supports the motor theory of speech perception. A critical problem in vocal imitation is how to generate speech sounds produced by adults, whose vocal tracts have physical properties (i.e., articulatory motions) differing from those of infants' vocal tracts. To solve this problem, a model based on the motor theory of speech perception, was constructed. Applying this model enables the vocal imitation system to estimate articulatory motions for unexperienced speech sounds that have not actually been generated by the system. The system was implemented by using Recurrent Neural Network with Parametric Bias (RNNPB) and a physical vocal tract model, called Maeda model. Experimental results demonstrated that the system was sufficiently robust with respect to individual differences in speech sounds and could imitate unexperienced vowel sounds.

UR - http://www.scopus.com/inward/record.url?scp=54049109678&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=54049109678&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-69162-4_24

DO - 10.1007/978-3-540-69162-4_24

M3 - Conference contribution

AN - SCOPUS:54049109678

SN - 3540691596

SN - 9783540691594

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 222

EP - 232

BT - Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers

T2 - 14th International Conference on Neural Information Processing, ICONIP 2007

Y2 - 13 November 2007 through 16 November 2007

ER -

Vowel imitation using vocal tract model and recurrent neural network

Abstract

Publication series

Conference

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this