Many-to-Many Voice Conversion with Out-of-Dataset Speaker Support
arXiv:1905.02525
Abstract
We present a Cycle-GAN based many-to-many voice conversion method that can convert between speakers that are not in the training set. This property is enabled through speaker embeddings generated by a neural network that is jointly trained with the Cycle-GAN. In contrast to prior work in this domain, our method enables conversion between an out-of-dataset speaker and a target speaker in either direction and does not require re-training. Out-of-dataset speaker conversion quality is evaluated using an independently trained speaker identification model, and shows good style conversion characteristics for previously unheard speakers. Subjective tests on human listeners show style conversion quality for in-dataset speakers is comparable to the state-of-the-art baseline model.
arXiv commentSubmitted to Interspeech 2019
References in corpus (6)
- Instance Normalization: The Missing Ingredient for Fast Stylization
- Parallel-Data-Free Voice Conversion Using Cycle-Consistent Adversarial Networks
- Voice Conversion from Unaligned Corpora using Variational Autoencoding Wasserstein Generative Adversarial Networks
- Checkerboard artifact free sub-pixel convolution: A note on sub-pixel convolution, resize convolution and convolution resize
- ACVAE-VC: Non-parallel many-to-many voice conversion with auxiliary classifier variational autoencoder
- StarGAN-VC: Non-parallel many-to-many voice conversion with star generative adversarial networks