Emotional Voice Conversion using Multitask Learning with Text-to-speech
Authors:
- Tae-Ho KimTKTae-Ho Kim
- Sungjae ChoSCSungjae Cho
- Sejik ParkSPSejik Park
- Soo-Young LeeSLSoo-Young Lee
Preprint
Abstract
Voice conversion (VC) is a task to transform a person's voice to different style while conserving linguistic contents. Previous state-of-the-art on VC is based on sequence-to-sequence (seq2seq) model, which could mislead linguistic information. There was an attempt to overcome it by using textual supervision, it requires explicit alignment which loses the benefit of using seq2seq model. In this paper, a voice converter using multitask learning with text-to-speech (TTS) is presented. The embedding space of seq2seq-based TTS has abundant information on the text. The role of the decoder of TTS is to convert embedding space to speech, which is same to VC. In the proposed model, the whole network is trained to minimize loss of VC and TTS. VC is expected to capture more linguistic information and to preserve training stability by multitask learning. Experiments of VC were performed on a male Korean emotional text-speech dataset, and it is shown that multitask learning is helpful to keep linguistic contents in VC.Keywords
All Related Versions
This publication has 15 references indexed in Scilit:
- Emotional Voice Conversion Using Dual Supervised Adversarial Networks With Continuous Wavelet Transform F0 FeaturesIEEE/ACM Transactions on Audio, Speech, and Language Processing, 2019
- Improving Sequence-to-sequence Voice Conversion by Adding Text-supervision2019
- Sequence-to-Sequence Acoustic Modeling for Voice ConversionIEEE/ACM Transactions on Audio, Speech, and Language Processing, 2019
- Voice conversion for emotional speech: Rule-based synthesis with degree of emotion controllable in dimensional spaceSpeech Communication, 2018
- Natural TTS Synthesis by Conditioning Wavenet on MEL Spectrogram Predictions2018
- Emotional voice conversion using neural networks with arbitrary scales F0 based on wavelet transformEURASIP Journal on Audio, Speech, and Music Processing, 2017
- Voice conversion using deep Bidirectional Long Short-Term Memory based Recurrent Neural Networks2015
- Spectral Mapping Using Artificial Neural Networks for Voice ConversionIEEE Transactions on Audio, Speech, and Language Processing, 2010
- Voice Conversion Based on Maximum-Likelihood Estimation of Spectral Parameter TrajectoryIEEE Transactions on Audio, Speech, and Language Processing, 2007
- Korean large vocabulary continuous speech recognition with morpheme-based recognition unitsSpeech Communication, 2002