arrow
返回

MASS: Multi-task anthropomorphic speech synthesis framework

delete2021-11-01
delete3
delete
OA
AI
陈
陈晋音 (Jinyin Chen) *
L
Linhui Ye
Z
Zhaoyan Ming
DOI:10.1016/j.csl.2021.101243delete
delete原文链接
delete分享
delete收藏
查看原文
摘要

摘要

En 中文
Text-to-Speech (TTS) synthesis plays an important role in human-computer interaction. Currently, most TTS technologies focus on the naturalness of speech, namely, making the speeches sound like humans. However, the key tasks of the expression of emotion and the speaker identity are ignored, which limits the application scenarios of TTS synthesis technology. To make the synthesized speech more realistic and expand the application scenarios, we propose a multi-task anthropomorphic speech synthesis framework (MASS), which can synthesize speeches from text with specified emotion and speaker identity. The MASS framework consists of a base TTS module and two novel voice conversion modules: the emotional voice conversion module and the speaker voice conversion module. We propose deep emotion voice conversion model (DEVC) and deep speaker voice conversion model (DSVC) based on convolution residual networks. It solves the problem of feature loss during voice conversion. The model trainings are independent of parallel datasets, and are capable of many-to many voice conversion. In the emotional voice conversion, speaker voice conversion experiments, as well as the multi-task speech synthesis experiments, experimental results show DEVC and DSVC convert speech effectively. The quantitative and qualitative evaluation results of multi-task speech synthesis experiments show MASS can effectively synthesis speech with specified text, emotion and speaker identity . (c) 2021 Elsevier Ltd. All rights reserved.
Keyword:
Text-to-speech
Emotional voice conversion
Speaker voice conversion
Convolution residual network
AI总结

AI总结

对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。

期刊

C
Computer Speech and Language
IF:
3.4
论文数:
1.5K
被引数:
2.6K

机构

Z
zhejiang university of technology
学者数:
3.3W
论文数: 2.0W
被引数: 22
Z
zhejiang university
学者数:
17.7W
论文数: 12.1W
被引数: 152
引用论文

引用论文

Characterization of the colistin (polymyxin E1 and E2) biosynthetic gene cluster
err2015-01-22
err0
PREAI
errFatoumata Tambadou; Thibault Caradec; Anne-Laure Gagez; Antoine Bonnet; Valérie Sopéna; Nicolas Bridiau; Valérie Thiéry; Sandrine Didelot; Cyrille Barthélémy; Romain Chevrot
err分享
err收藏
Fragment Screening and Fast Micromolar Detection on a Benchtop NMR Spectrometer Boosted by Photoinduced Hyperpolarization
err2023-08-24
err0
errOAAI
errGabriela R. Stadler; Takuya F. Segawa; Matthias Bütikofer; Venita Decker; Sandra Loss; Barbara Czarniecki; Felix Torres; Roland Riek
err分享
err收藏
Stem and Somatic Cell Monotherapy for the Treatment of Diabetic Foot Ulcers: Review of Clinical Studies and Mechanisms of Action
err2022-04-27
err0
PREAI
errO. A. Krasilnikova; D. S. Baranovskii; A. V. Lyundup; P. V. Shegay; A. D. Kaprin; I. D. Klabukov
err分享
err收藏
Quantitative Contributions of the Muscles of the Tongue, Floor-of-Mouth, Jaw, and Velum to Tongue-to-Palate Pressure Generation
err2008-08-01
err0
PREAI
errPhyllis M. Palmer; Debra M. Jaffe; Timothy M. McCulloch; Eileen M. Finnegan; Douglas J. Van Daele; Erich S. Luschei
err分享
err收藏
学者 查看更多内容