Services

AI Voice Models (RVC)

RVC stands for Retrieval-based Voice Conversion. We train a custom model on a voice, then use it to turn a spoken or sung performance into that voice, keeping the timing and emotion of the original. A character can sound the same in every scene, and one performer can voice a whole cast.

Custom AI voice models with RVC

RVC (Retrieval-based Voice Conversion) is a form of AI voice cloning. We train a custom model on recordings of a voice. A performer then speaks or sings the lines, and the model turns that performance into the trained voice, keeping the timing, phrasing and emotion of the original.

Because the performance comes from a person, the result sounds acted rather than read by a machine.

What custom voice models are used for

Consent comes first

We train a model on a real person's voice only with that person's written consent. You own the voices we create for your production, and the third-party tools we use are licensed for commercial use.

Voice models pair with our influencer, animated character and music video productions, so the voice stays as consistent as the picture.

Voice model questions

What is RVC voice model training?

RVC stands for Retrieval-based Voice Conversion. We train a custom model on a voice and use it to turn a spoken or sung performance into that voice, keeping the timing and emotion of the original. A character can sound the same in every scene, and one performer can voice a whole cast. We train a model on a real person's voice only with their written consent.

Do you use real people's faces or voices?

Only with written consent. We use a real person's face or voice, including for a custom voice model, only with that person's written consent.

Contact

One second or a feature film.
Let's talk about your project.