How much voice data is required to create a Local Voice Clone?

The minimum amount of voice data required is only 2 minutes. Additional data improves the results, but the improvement tends to saturate somewhere between 20 and 60 min of recordings, depending on the quality of the recordings and the distinctiveness of the speaker. We strongly suggest to use the maximum ammount of clean data you can find, in order to be used in project with post-production quality.