Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

AttS2S-VC: Sequence-to-Sequence Voice Conversion with Attention and Context Preservation Mechanisms

Abstract: 

This paper describes a method based on a sequence-to-sequence learning (Seq2Seq) with attention and context preservation mechanism for voice conversion (VC) tasks. Seq2Seq has been outstanding at numerous tasks involving sequence modeling such as speech synthesis and recognition, machine translation, and image captioning. In contrast to current VC techniques, our method 1) stabilizes and accelerates the training procedure by considering guided attention and proposed context preservation losses, 2) allows not only spectral envelopes but also fundamental frequency contours and durations of speech to be converted, 3) requires no context information such as phoneme labels, and 4) requires no time-aligned source and target speech data in advance. In our experiment, the proposed VC framework can be trained in only one day, using only one GPU of an NVIDIA Tesla K80, while the quality of the synthesized speech is higher than that of speech converted by Gaussian mixture model-based VC and is comparable to that of speech generated by recurrent neural network-based text-to-speech synthesis, which can be regarded as an upper limit on VC performance.

up
0 users have voted:

Paper Details

Authors:
Submitted On:
15 May 2019 - 7:03am
Short Link:
Type:
Poster
Event:
Presenter's Name:
Kou Tanaka
Paper Code:
2505
Document Year:
2019
Cite

Document Files

2019_05_ICASSP_KouTanaka.pdf

(6)

Subscribe

[1] , "AttS2S-VC: Sequence-to-Sequence Voice Conversion with Attention and Context Preservation Mechanisms", IEEE SigPort, 2019. [Online]. Available: http://sigport.org/4522. Accessed: May. 23, 2019.
@article{4522-19,
url = {http://sigport.org/4522},
author = { },
publisher = {IEEE SigPort},
title = {AttS2S-VC: Sequence-to-Sequence Voice Conversion with Attention and Context Preservation Mechanisms},
year = {2019} }
TY - EJOUR
T1 - AttS2S-VC: Sequence-to-Sequence Voice Conversion with Attention and Context Preservation Mechanisms
AU -
PY - 2019
PB - IEEE SigPort
UR - http://sigport.org/4522
ER -
. (2019). AttS2S-VC: Sequence-to-Sequence Voice Conversion with Attention and Context Preservation Mechanisms. IEEE SigPort. http://sigport.org/4522
, 2019. AttS2S-VC: Sequence-to-Sequence Voice Conversion with Attention and Context Preservation Mechanisms. Available at: http://sigport.org/4522.
. (2019). "AttS2S-VC: Sequence-to-Sequence Voice Conversion with Attention and Context Preservation Mechanisms." Web.
1. . AttS2S-VC: Sequence-to-Sequence Voice Conversion with Attention and Context Preservation Mechanisms [Internet]. IEEE SigPort; 2019. Available from : http://sigport.org/4522