Paper Abstract and Keywords |
Presentation |
2013-07-25 17:30
Voice conversion using piecewise linear transformation with long-span features and state discrimination of the target speaker Jun Ikeshima, Masayuki Suzuki, Daisuke Saito, Nobuaki Minematsu, Keikichi Hirose (Univ. of Tokyo) SP2013-56 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
A method has been proposed for voice conversion. It is based on piecewise linear transformation with long time-span features, and state discrimination of the target speaker. The piecewise linear transformation consists of two methods in order to realize nonlinear transformation between source and target speakers as a set of linear transformations; dividing acoustic feature space into several regions and approximating a linear transformation for each region. This idea has been implemented as ``soft'' space division using GMM (Gaussian Mixture Model) and successfully applied to voice conversion. In the original/conventional method, space division is conducted using feature vectors of source speaker or joint feature vectors of source and target speakers. However, resulted division is not in good much with feature space of target speaker, when there are large differences between distributions of feature spaces of source and target speakers. In order to solve this situation, a new method of space division by features of target speaker is developed. Since features of target speaker are not available during conversion, ”affiliation probability” is estimated using discriminative modeling scheme. Also features of time-adjacent frames are added to the feature vectors (of discrimination and transformation processes), corresponding to the fact that the long term features are important for voice conversion similar to the case of speech recognition. The proposed method was shown to be effective for voice conversion through objective and subjective evaluations. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
voice conversion / piecewise linear transformation / discrimination model / long span feature / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 113, no. 161, SP2013-56, pp. 19-24, July 2013. |
Paper # |
SP2013-56 |
Date of Issue |
2013-07-18 (SP) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
SP2013-56 |
Conference Information |
Committee |
SP IPSJ-SLP |
Conference Date |
2013-07-25 - 2013-07-27 |
Place (in Japanese) |
(See Japanese page) |
Place (in English) |
Soho (togatta spa) |
Topics (in Japanese) |
(See Japanese page) |
Topics (in English) |
Speech recognition and understanding, dialog system, etc. |
Paper Information |
Registration To |
SP |
Conference Code |
2013-07-SP-SLP |
Language |
Japanese |
Title (in Japanese) |
(See Japanese page) |
Sub Title (in Japanese) |
(See Japanese page) |
Title (in English) |
Voice conversion using piecewise linear transformation with long-span features and state discrimination of the target speaker |
Sub Title (in English) |
|
Keyword(1) |
voice conversion |
Keyword(2) |
piecewise linear transformation |
Keyword(3) |
discrimination model |
Keyword(4) |
long span feature |
Keyword(5) |
|
Keyword(6) |
|
Keyword(7) |
|
Keyword(8) |
|
1st Author's Name |
Jun Ikeshima |
1st Author's Affiliation |
The University of Tokyo (Univ. of Tokyo) |
2nd Author's Name |
Masayuki Suzuki |
2nd Author's Affiliation |
The University of Tokyo (Univ. of Tokyo) |
3rd Author's Name |
Daisuke Saito |
3rd Author's Affiliation |
The University of Tokyo (Univ. of Tokyo) |
4th Author's Name |
Nobuaki Minematsu |
4th Author's Affiliation |
The University of Tokyo (Univ. of Tokyo) |
5th Author's Name |
Keikichi Hirose |
5th Author's Affiliation |
The University of Tokyo (Univ. of Tokyo) |
6th Author's Name |
|
6th Author's Affiliation |
() |
7th Author's Name |
|
7th Author's Affiliation |
() |
8th Author's Name |
|
8th Author's Affiliation |
() |
9th Author's Name |
|
9th Author's Affiliation |
() |
10th Author's Name |
|
10th Author's Affiliation |
() |
11th Author's Name |
|
11th Author's Affiliation |
() |
12th Author's Name |
|
12th Author's Affiliation |
() |
13th Author's Name |
|
13th Author's Affiliation |
() |
14th Author's Name |
|
14th Author's Affiliation |
() |
15th Author's Name |
|
15th Author's Affiliation |
() |
16th Author's Name |
|
16th Author's Affiliation |
() |
17th Author's Name |
|
17th Author's Affiliation |
() |
18th Author's Name |
|
18th Author's Affiliation |
() |
19th Author's Name |
|
19th Author's Affiliation |
() |
20th Author's Name |
|
20th Author's Affiliation |
() |
Speaker |
Author-1 |
Date Time |
2013-07-25 17:30:00 |
Presentation Time |
30 minutes |
Registration for |
SP |
Paper # |
SP2013-56 |
Volume (vol) |
vol.113 |
Number (no) |
no.161 |
Page |
pp.19-24 |
#Pages |
6 |
Date of Issue |
2013-07-18 (SP) |
|