Voice conversion through vector quantization

Masanobu Abe, Satoshi Nakamura, Kiyohiro Shikano, Hisao Kuwabara

Research output: Contribution to journalArticle

56 Citations (Scopus)

Abstract

A new voice conversion technique through vector quantization and spectrum mapping is proposed. This technique is based on mapping codebooks which represent the correspondence between different speakers’ codebooks. The mapping codebooks for spectrum parameters, power values, and pitch frequencies are separately generated using training utterances. This technique makes it possible to precisely control voice individuality. The performance of this technique is confirmed by spectrum distortion and pitch frequency difference. To evaluate the overall performance of this technique, listening tests are carried out on two kinds of voice conversions: one between male and female speakers, the other between male speakers. In the male-to-female conversion experiment, all converted utterances are judged as female, and in the male-to-male conversion, 57 % of them are identified as the target speaker.

Original languageEnglish
Pages (from-to)71-76
Number of pages6
JournalJournal of the Acoustical Society of Japan (E)
Volume11
Issue number2
DOIs
Publication statusPublished - 1990
Externally publishedYes

Fingerprint

vector quantization
voice control
education

ASJC Scopus subject areas

  • Acoustics and Ultrasonics

Cite this

Voice conversion through vector quantization. / Abe, Masanobu; Nakamura, Satoshi; Shikano, Kiyohiro; Kuwabara, Hisao.

In: Journal of the Acoustical Society of Japan (E), Vol. 11, No. 2, 1990, p. 71-76.

Research output: Contribution to journalArticle

Abe, Masanobu ; Nakamura, Satoshi ; Shikano, Kiyohiro ; Kuwabara, Hisao. / Voice conversion through vector quantization. In: Journal of the Acoustical Society of Japan (E). 1990 ; Vol. 11, No. 2. pp. 71-76.
@article{90ed28e42a144bf2a9733a590ab58765,
title = "Voice conversion through vector quantization",
abstract = "A new voice conversion technique through vector quantization and spectrum mapping is proposed. This technique is based on mapping codebooks which represent the correspondence between different speakers’ codebooks. The mapping codebooks for spectrum parameters, power values, and pitch frequencies are separately generated using training utterances. This technique makes it possible to precisely control voice individuality. The performance of this technique is confirmed by spectrum distortion and pitch frequency difference. To evaluate the overall performance of this technique, listening tests are carried out on two kinds of voice conversions: one between male and female speakers, the other between male speakers. In the male-to-female conversion experiment, all converted utterances are judged as female, and in the male-to-male conversion, 57 {\%} of them are identified as the target speaker.",
author = "Masanobu Abe and Satoshi Nakamura and Kiyohiro Shikano and Hisao Kuwabara",
year = "1990",
doi = "10.1250/ast.11.71",
language = "English",
volume = "11",
pages = "71--76",
journal = "Acoustical Science and Technology",
issn = "1346-3969",
publisher = "Acoustical Society of Japan",
number = "2",

}

TY - JOUR

T1 - Voice conversion through vector quantization

AU - Abe, Masanobu

AU - Nakamura, Satoshi

AU - Shikano, Kiyohiro

AU - Kuwabara, Hisao

PY - 1990

Y1 - 1990

N2 - A new voice conversion technique through vector quantization and spectrum mapping is proposed. This technique is based on mapping codebooks which represent the correspondence between different speakers’ codebooks. The mapping codebooks for spectrum parameters, power values, and pitch frequencies are separately generated using training utterances. This technique makes it possible to precisely control voice individuality. The performance of this technique is confirmed by spectrum distortion and pitch frequency difference. To evaluate the overall performance of this technique, listening tests are carried out on two kinds of voice conversions: one between male and female speakers, the other between male speakers. In the male-to-female conversion experiment, all converted utterances are judged as female, and in the male-to-male conversion, 57 % of them are identified as the target speaker.

AB - A new voice conversion technique through vector quantization and spectrum mapping is proposed. This technique is based on mapping codebooks which represent the correspondence between different speakers’ codebooks. The mapping codebooks for spectrum parameters, power values, and pitch frequencies are separately generated using training utterances. This technique makes it possible to precisely control voice individuality. The performance of this technique is confirmed by spectrum distortion and pitch frequency difference. To evaluate the overall performance of this technique, listening tests are carried out on two kinds of voice conversions: one between male and female speakers, the other between male speakers. In the male-to-female conversion experiment, all converted utterances are judged as female, and in the male-to-male conversion, 57 % of them are identified as the target speaker.

UR - http://www.scopus.com/inward/record.url?scp=85004448479&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85004448479&partnerID=8YFLogxK

U2 - 10.1250/ast.11.71

DO - 10.1250/ast.11.71

M3 - Article

VL - 11

SP - 71

EP - 76

JO - Acoustical Science and Technology

JF - Acoustical Science and Technology

SN - 1346-3969

IS - 2

ER -