The AS-NU System for the M2VoC Challenge
This addresses voice cloning for low-resource scenarios, but it is incremental as it adapts existing methods to specific challenge constraints.
The paper tackled voice cloning with limited data in two challenge tracks, achieving similar quality but lower similarity scores when using only 5 utterances compared to 100 utterances.
This paper describes the AS-NU systems for two tracks in MultiSpeaker Multi-Style Voice Cloning Challenge (M2VoC). The first track focuses on using a small number of 100 target utterances for voice cloning, while the second track focuses on using only 5 target utterances for voice cloning. Due to the serious lack of data in the second track, we selected the speaker most similar to the target speaker from the training data of the TTS system, and used the speaker's utterances and the given 5 target utterances to fine-tune our model. The evaluation results show that our systems on the two tracks perform similarly in terms of quality, but there is still a clear gap between the similarity score of the second track and the similarity score of the first track.