An instant clone needs one to three minutes of clean mono audio and is good enough for a scratch read. A professionally trained clone needs thirty minutes at minimum, two to three hours for the strongest result, and holds emotional range far better, which is the part that decides whether a listener stays.
Consent is not a formality here. A signed release for the voice sits in the folder before anything renders, and that applies to a founder’s own voice as much as to anyone else’s.
How much audio do you need to clone a voice?
One to three minutes for an instant clone, thirty minutes to three hours for a trained one. The longer version is noticeably better at emotion, pacing and emphasis.
Is voice cloning legal for advertising?
With the speaker’s written consent and appropriate disclosure, generally yes. Without consent it engages personality rights, and increasingly statute. We do not render without the release.