Paper Abstract and Keywords |
Presentation |
2017-01-21 16:35
Simultaneous modeling of acoustic feature sequences and its temporal structures for DNN-based speech synthesis Kei Hashimoto, Keiichiro Oura, Yoshihiko Nankaku, Keiichi Tokuda (Nagoya Inst. of Tech.) SP2016-76 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
In statistical parametric speech synthesis, a hidden Markov model (HMM) is widely used as an acoustic model. Recently, deep neural networks (DNNs) are used as acoustic models and DNN-based systems show better performance than HMM-based ones. In the conventional DNN-based speech synthesis, acoustic features are modeled by DNNs. However, duration information is given from other external duration predictors and temporal structures of speech are not considered in training of DNNs. In this paper, we propose a novel DNN-based acoustic model that outputs parameters of hidden semi-Markov models (HSMMs), which is an HMM with explicit duration models. The proposed method can simultaneously model acoustic feature sequences and its temporal structures in a unified framework. Experimental results show that the proposed method can generate natural-sounding speech as achieving a significant computational cost reduction. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
speech synthesis / deep neural network / acoustic model / duration model / hidden semi-Markov model / / / |
Reference Info. |
IEICE Tech. Rep., vol. 116, no. 414, SP2016-76, pp. 71-76, Jan. 2017. |
Paper # |
SP2016-76 |
Date of Issue |
2017-01-14 (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 |
SP2016-76 |
Conference Information |
Committee |
SP |
Conference Date |
2017-01-21 - 2017-01-21 |
Place (in Japanese) |
(See Japanese page) |
Place (in English) |
The University of Tokyo |
Topics (in Japanese) |
(See Japanese page) |
Topics (in English) |
Synthesis, Generation, Prosody, etc. |
Paper Information |
Registration To |
SP |
Conference Code |
2017-01-SP |
Language |
Japanese |
Title (in Japanese) |
(See Japanese page) |
Sub Title (in Japanese) |
(See Japanese page) |
Title (in English) |
Simultaneous modeling of acoustic feature sequences and its temporal structures for DNN-based speech synthesis |
Sub Title (in English) |
|
Keyword(1) |
speech synthesis |
Keyword(2) |
deep neural network |
Keyword(3) |
acoustic model |
Keyword(4) |
duration model |
Keyword(5) |
hidden semi-Markov model |
Keyword(6) |
|
Keyword(7) |
|
Keyword(8) |
|
1st Author's Name |
Kei Hashimoto |
1st Author's Affiliation |
Nagoya Institute of Technology (Nagoya Inst. of Tech.) |
2nd Author's Name |
Keiichiro Oura |
2nd Author's Affiliation |
Nagoya Institute of Technology (Nagoya Inst. of Tech.) |
3rd Author's Name |
Yoshihiko Nankaku |
3rd Author's Affiliation |
Nagoya Institute of Technology (Nagoya Inst. of Tech.) |
4th Author's Name |
Keiichi Tokuda |
4th Author's Affiliation |
Nagoya Institute of Technology (Nagoya Inst. of Tech.) |
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 |
2017-01-21 16:35:00 |
Presentation Time |
25 minutes |
Registration for |
SP |
Paper # |
SP2016-76 |
Volume (vol) |
vol.116 |
Number (no) |
no.414 |
Page |
pp.71-76 |
#Pages |
6 |
Date of Issue |
2017-01-14 (SP) |
|