GR0: Self-Supervised Global Representation Learning for Zero-Shot Voice Conversion | IEEE Conference Publication | IEEE Xplore