Voice Cloning & Consent

Zero-shot cloning: a few seconds of reference audio → a speaker embedding → any text spoken in that voice, no training. Few-shot fine-tuning: minutes of audio → a dedicated model adaptation → higher fidelity and range.

What cloning captures: timbre, accent, baseline prosody — the recognizable 'who'. What it misses: the actor — emotional range, character work, the trained instrument of a professional voice. Cloning is excellent at sounding like someone; it's mediocre at being them.