Meta-Voice: Fast few-shot style transfer for expressive voice cloning using meta learning
This addresses the problem of rapid voice cloning for real-world applications, though it is incremental as it builds on existing meta-learning and disentanglement techniques.
The paper tackled the challenging task of fast few-shot style transfer for voice cloning in text-to-speech synthesis by using meta learning, achieving voice cloning with only 5 samples (about 12 seconds of speech) and 100 adaptation steps.
The task of few-shot style transfer for voice cloning in text-to-speech (TTS) synthesis aims at transferring speaking styles of an arbitrary source speaker to a target speaker's voice using very limited amount of neutral data. This is a very challenging task since the learning algorithm needs to deal with few-shot voice cloning and speaker-prosody disentanglement at the same time. Accelerating the adaptation process for a new target speaker is of importance in real-world applications, but even more challenging. In this paper, we approach to the hard fast few-shot style transfer for voice cloning task using meta learning. We investigate the model-agnostic meta-learning (MAML) algorithm and meta-transfer a pre-trained multi-speaker and multi-prosody base TTS model to be highly sensitive for adaptation with few samples. Domain adversarial training mechanism and orthogonal constraint are adopted to disentangle speaker and prosody representations for effective cross-speaker style transfer. Experimental results show that the proposed approach is able to conduct fast voice cloning using only 5 samples (around 12 second speech data) from a target speaker, with only 100 adaptation steps. Audio samples are available online.