Noise-robust voice conversion with domain adversarial training
dhqadg.github.io