How long is the training data suitable? The more, the better？

[ZLZ-Nancy]

I have used 170 corpus training, The effect is still good. Is it necessary to add sample data to get a bigger effect？

[unilight]

In theory, a bigger training dataset leads to a better performance (generalization ability), but I think for GMM-VC about 50-80 utterances should be enough, i.e. the improvement can be marginal if you add in more data after you dataset reach this size.