Information: Join today and make your research activities more affordable! Technical workshop participation fees and annual registration fees are available at member rates.
Notice: [Important] Announcement of Changes to Registration Fee Payment and Manuscript Upload Procedures for IEICE Technical Meetings
IEICE Technical Committee Submission System
Conference Paper's Information		 Online Proceedings
[Sign in]
Tech. Rep. Archives
 Go Top Page Go Previous   [Japanese] / [English] 

Paper Abstract and Keywords
Presentation 2023-02-28 09:30
MS-FC-HiFiGAN : Fast Neural Waveform Generation Model With Learnable Lightweight Upsampling
Haruki Yamashita (Kobe Univ/NICT), Takuma Okamoto (NICT), Ryoichi Takashima, Tetsuya Takiguchi (Kobe Univ), Tomoki Toda (Nagoya Univ/NICT), Hisashi Kawai (NICT) EA2022-76 SIP2022-120 SP2022-40
Abstract (in Japanese) (See Japanese page) 
(in English) In recent years, in text-to-speech synthesis, it is required to improve the inference speed while keeping the quality.
Multi-stream(MS) iSTFT-HiFiGAN was proposed as a high-speed model of HiFi-GAN, a vocoder capable of inferring waveforms on single CPU.
In the TTS task using VITS, although there was some deterioration in sound quality, the speed was increased by about 4 times.
In this paper, we propose a MS-FC-HiFi-GAN in which the inverse short-time Fourier transform (iSTFT) part is changed to trainable fully connected layer for the purpose of improving the synthesis quality of the MS-iSTFT-HiFiGAN.
As for the inference speed, RTF was 0.15 on 1 CPU, which is the same as MS-iSTFT-HiFiGAN.
Synthesis quality was inferior to that of MS-iSTFT-HiFiGAN in TTS task, but was superior to thatin analysis/synthesis task.
Keyword (in Japanese) (See Japanese page) 
(in English) speech synthesis / Neural Vocoder / HiFi-GAN / Text-to-Speech / Analysis Synthesis / / /  
Reference Info. IEICE Tech. Rep., vol. 122, no. 389, SP2022-40, pp. 7-12, Feb. 2023.
Paper # SP2022-40 
Date of Issue 2023-02-21 (EA, SIP, SP) 
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 EA2022-76 SIP2022-120 SP2022-40

Conference Information
Committee SP IPSJ-SLP EA SIP  
Conference Date 2023-02-28 - 2023-03-01 
Place (in Japanese) (See Japanese page) 
Place (in English)  
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To SP 
Conference Code 2023-02-SP-SLP-EA-SIP 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) MS-FC-HiFiGAN : Fast Neural Waveform Generation Model With Learnable Lightweight Upsampling 
Sub Title (in English)  
Keyword(1) speech synthesis  
Keyword(2) Neural Vocoder  
Keyword(3) HiFi-GAN  
Keyword(4) Text-to-Speech  
Keyword(5) Analysis Synthesis  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Haruki Yamashita  
1st Author's Affiliation Kobe University/National Institute of Information and Communications Technology (Kobe Univ/NICT)
2nd Author's Name Takuma Okamoto  
2nd Author's Affiliation National Institute of Information and Communications Technology (NICT)
3rd Author's Name Ryoichi Takashima  
3rd Author's Affiliation Kobe University (Kobe Univ)
4th Author's Name Tetsuya Takiguchi  
4th Author's Affiliation Kobe University (Kobe Univ)
5th Author's Name Tomoki Toda  
5th Author's Affiliation Nagoya University/National Institute of Information and Communications Technology (Nagoya Univ/NICT)
6th Author's Name Hisashi Kawai  
6th Author's Affiliation National Institute of Information and Communications Technology (NICT)
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 ()
21st Author's Name  
21st Author's Affiliation ()
22nd Author's Name  
22nd Author's Affiliation ()
23rd Author's Name  
23rd Author's Affiliation ()
24th Author's Name  
24th Author's Affiliation ()
25th Author's Name  
25th Author's Affiliation ()
26th Author's Name / /
26th Author's Affiliation ()
()
27th Author's Name / /
27th Author's Affiliation ()
()
28th Author's Name / /
28th Author's Affiliation ()
()
29th Author's Name / /
29th Author's Affiliation ()
()
30th Author's Name / /
30th Author's Affiliation ()
()
31st Author's Name / /
31st Author's Affiliation ()
()
32nd Author's Name / /
32nd Author's Affiliation ()
()
33rd Author's Name / /
33rd Author's Affiliation ()
()
34th Author's Name / /
34th Author's Affiliation ()
()
35th Author's Name / /
35th Author's Affiliation ()
()
36th Author's Name / /
36th Author's Affiliation ()
()
Speaker Author-1 
Date Time 2023-02-28 09:30:00 
Presentation Time 20 minutes 
Registration for SP 
Paper # EA2022-76, SIP2022-120, SP2022-40 
Volume (vol) vol.122 
Number (no) no.387(EA), no.388(SIP), no.389(SP) 
Page pp.7-12 
#Pages 6 
Date of Issue 2023-02-21 (EA, SIP, SP) 

[Return to Top Page]

[Return to IEICE Web Page]
The Institute of Electronics, Information and Communication Engineers (IEICE), Japan