My Publications
I started working on voice conversion since 2016. My goal is to make it easier and more readily applicable to realistic circumstances.
Voice Conversion from Unaligned Corpora using Variational Autoencoding Wasserstein Generative Adversarial Networks
Chin-Chen Hsu, Hsin-Te Hwang, Yi-Chiao Wu, Yu Tsao, and Hsin-Min Wang
Proceedings of INTERSPEECH 2017 (in press)
We propose a non-parallel VC framework with a Wasserstein generative adversarial network (W-GAN) that explicitly takes a VC-related objective into account. Experimental results corroborated the capability of our framework for building a VC system from unaligned data, and demonstrated improved conversion quality.
Voice Conversion from Non-parallel Corpora Using Variational Auto-encoder
Chin-Chen Hsu, Hsin-Te Hwang, Yi-Chiao Wu, Yu Tsao, and Hsin-Min Wang
Proceedings of APSIPA, 2016
By means of deep generative models, we are able to liberate the task of voice conversion from the need of parallel corpora (i.e. strict supervision), making the training process easier.
Dictionary Update for NMF-based Voice Conversion Using an Encoder-Decoder Network
Chin-Chen Hsu, Hsin-Te Hwang, Yi-Chiao Wu, Yu Tsao, and Hsin-Min Wang
Proceedings of ISCSLP 2016.
We made dictionary update easier through reformulating it as an auto-encoder based on deep neural networks. It also significantly improves voice quality of the synthetic speech.
A Post-filtering Approach Based on Locally Linear Embedding Difference Compensation for Speech Enhancement
Yi-Chiao Wu, Hsin-Te Hwang, Syu-Siang Wang, Chin-Cheng Hsu, Yu Tsao, Hsin-Min Wang
Proceedings of INTERSPEECH 2017 (in press)
Locally Linear Embedding for Exemplar-Based Spectral Conversion
Yi-Chiao Wu, Hsin-Te Hwang, Chin-Cheng Hsu, Yu Tsao, Hsin-Min Wang
Proceedings of INTERSPEECH, 2016
We made dictionary update easier through reformulating it as an auto-encoder based on deep neural networks. It also significantly improves voice quality of the synthetic speech.
Robust Emotion Recognition by Spectro-Temporal Modulation Statistic Features
Tai-Shi Chi, Lan-Ying Yeh and Chin-Cheng Hsu
Journal of Ambient Intelligence and Humanized Computing, 2011
On Robustness of Spectro-Temporal Modulation Features in an Emotion Recognition Framework
Author: Chin-Cheng Hsu
Advisor: Tai-Shi Chi
Master's Thesis
Patent
Text-to-speech method and multi-lingual speech synthesizer using the method
Authors: Hsun-Fu Liu, Abhishek Pandey, Chin-Cheng Hsu
Patent ID: US20170047060A1