Picture for Kun Song

Kun Song

Multi-Speaker Expressive Speech Synthesis via Multiple Factors Decoupling

Nov 19, 2022
Xinfa Zhu, Yi Lei, Kun Song, Yongmao Zhang, Tao Li, Lei Xie

Figure 1 for Multi-Speaker Expressive Speech Synthesis via Multiple Factors Decoupling

Figure 2 for Multi-Speaker Expressive Speech Synthesis via Multiple Factors Decoupling

Figure 3 for Multi-Speaker Expressive Speech Synthesis via Multiple Factors Decoupling

Figure 4 for Multi-Speaker Expressive Speech Synthesis via Multiple Factors Decoupling

* Submitted to ICASSP2023

Access Paper or Ask Questions

Robust MelGAN: A robust universal neural vocoder for high-fidelity TTS

Nov 02, 2022
Kun Song, Jian Cong, Xinsheng Wang, Yongmao Zhang, Lei Xie, Ning Jiang, Haiying Wu

Figure 1 for Robust MelGAN: A robust universal neural vocoder for high-fidelity TTS

Figure 2 for Robust MelGAN: A robust universal neural vocoder for high-fidelity TTS

Figure 3 for Robust MelGAN: A robust universal neural vocoder for high-fidelity TTS

Figure 4 for Robust MelGAN: A robust universal neural vocoder for high-fidelity TTS

* Accepted by ISCSLP 2022

Access Paper or Ask Questions

DSPGAN: a GAN-based universal vocoder for high-fidelity TTS by time-frequency domain supervision from DSP

Nov 02, 2022
Kun Song, Yongmao Zhang, Yi Lei, Jian Cong, Hanzhao Li, Lei Xie, Gang He, Jinfeng Bai

Figure 1 for DSPGAN: a GAN-based universal vocoder for high-fidelity TTS by time-frequency domain supervision from DSP

Figure 2 for DSPGAN: a GAN-based universal vocoder for high-fidelity TTS by time-frequency domain supervision from DSP

Figure 3 for DSPGAN: a GAN-based universal vocoder for high-fidelity TTS by time-frequency domain supervision from DSP

Figure 4 for DSPGAN: a GAN-based universal vocoder for high-fidelity TTS by time-frequency domain supervision from DSP

* Submitted to ICASSP 2023

Access Paper or Ask Questions

Adaptive neighborhood Metric learning

Jan 20, 2022
Kun Song, Junwei Han, Gong Cheng, Jiwen Lu, Feiping Nie

* 14 pages, 6 figures

Access Paper or Ask Questions

The Multi-speaker Multi-style Voice Cloning Challenge 2021

Apr 05, 2021
Qicong Xie, Xiaohai Tian, Guanghou Liu, Kun Song, Lei Xie, Zhiyong Wu, Hai Li, Song Shi, Haizhou Li, Fen Hong, Hui Bu, Xin Xu

Figure 1 for The Multi-speaker Multi-style Voice Cloning Challenge 2021

Figure 2 for The Multi-speaker Multi-style Voice Cloning Challenge 2021

* has been accepted to ICASSP 2021

Access Paper or Ask Questions

Adaptive Nearest Neighbor: A General Framework for Distance Metric Learning

Nov 22, 2019
Kun Song

Figure 1 for Adaptive Nearest Neighbor: A General Framework for Distance Metric Learning

Figure 2 for Adaptive Nearest Neighbor: A General Framework for Distance Metric Learning

Figure 3 for Adaptive Nearest Neighbor: A General Framework for Distance Metric Learning

Figure 4 for Adaptive Nearest Neighbor: A General Framework for Distance Metric Learning

Access Paper or Ask Questions