Voice swap

A read with perfect timing in the wrong voice is one node from done. The Voice Changer keeps a performance — pacing, emphasis, emotion — and replaces the voice it’s delivered in.

The recipe

1. Get the performance — record the read yourself, or wire in any existing video or audio output on the canvas.
2. Wire it into Voice Changer and pick the target voice.
3. Run. Audio comes back converted; video comes back remuxed — same pixels, new voice — so you can feed a finished clip straight through without an extract-and-recombine dance.

Tips

• Perform the read the way it should sound — comedic timing, hesitation, emphasis all survive the swap. That’s the point of this over TTS: direction lives in the performance, not the prompt.
• Noisy source? Enable remove_background_noise before converting.
• Chain it after Character swap to recast a clip completely — new face, new voice, same shot.