Paper Abstract and Keywords |
Presentation |
2022-12-22 13:30
[Poster Presentation]
Quality Improvement of Children's Speech with Multiple Inputs of Speaker Vectors in a General Purpose Vocoder Satoshi Yoshida, Ken'ichi Furuya (Oita Univ.), Hideyuki Mizuno (SUS) EA2022-64 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Neural vocoders used in speech synthesis are capable of synthesizing high-quality speech that is indistinguishable from human speech. The objective of this research is to create a general-purpose vocoder that can synthesize high-quality speech for children and adults with a single speech synthesis model. In order to improve the quality of synthesized speech, we propose the introduction of a speaker vector that adds speaker information to the HiFi-GAN. We also consider the possibility of multiple inputs of the speaker vector to some or several layers of the generator of the HiFi-GAN. In order to confirm the effectiveness of the quality improvement, we compared the models trained with only acoustic features and each model with speaker vectors and evaluated the quality of the synthesized speech. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Vocoder / Speech Synthesis / Deep Learning / GAN / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 122, no. 324, EA2022-64, pp. 18-23, Dec. 2022. |
Paper # |
EA2022-64 |
Date of Issue |
2022-12-15 (EA) |
ISSN |
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 |
EA2022-64 |
Conference Information |
Committee |
EA US |
Conference Date |
2022-12-22 - 2022-12-23 |
Place (in Japanese) |
(See Japanese page) |
Place (in English) |
Satellite Campus Hiroshima |
Topics (in Japanese) |
(See Japanese page) |
Topics (in English) |
[Joint Meeting on Acoustics and Ultrasonics Subsociety] Engineering/Electro Acoustics, Ultrasonics, etc. |
Paper Information |
Registration To |
EA |
Conference Code |
2022-12-EA-US |
Language |
Japanese |
Title (in Japanese) |
(See Japanese page) |
Sub Title (in Japanese) |
(See Japanese page) |
Title (in English) |
Quality Improvement of Children's Speech with Multiple Inputs of Speaker Vectors in a General Purpose Vocoder |
Sub Title (in English) |
|
Keyword(1) |
Vocoder |
Keyword(2) |
Speech Synthesis |
Keyword(3) |
Deep Learning |
Keyword(4) |
GAN |
Keyword(5) |
|
Keyword(6) |
|
Keyword(7) |
|
Keyword(8) |
|
1st Author's Name |
Satoshi Yoshida |
1st Author's Affiliation |
Oita University (Oita Univ.) |
2nd Author's Name |
Ken'ichi Furuya |
2nd Author's Affiliation |
Oita University (Oita Univ.) |
3rd Author's Name |
Hideyuki Mizuno |
3rd Author's Affiliation |
Suwa University of Science (SUS) |
4th Author's Name |
|
4th Author's Affiliation |
() |
5th Author's Name |
|
5th Author's Affiliation |
() |
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 |
2022-12-22 13:30:00 |
Presentation Time |
120 minutes |
Registration for |
EA |
Paper # |
EA2022-64 |
Volume (vol) |
vol.122 |
Number (no) |
no.324 |
Page |
pp.18-23 |
#Pages |
6 |
Date of Issue |
2022-12-15 (EA) |
|