How to give each character in a dub their own voice
Give Vizard Agent a cast list rather than a language, and say that characters of the same gender need voices you can tell apart. It generates each character's lines into the existing timeline's windows so nothing drifts, mixes them into one dialogue track, then re-derives the captions from that mixed track rather than from the version they were timed to.
What do you type into Vizard Agent?
Name the characters and their relationships. "Use male and female voices" is a rule about gender; "the husband and the shop owner should not sound like the same man" is a rule about the scene, and only the second produces a usable dub. Vizard Agent can hear that there are several speakers, but not who is who.
Prompt
What does Vizard Agent actually do?
Dubbing is usually briefed as one job, this language into that one, so a single voice ends up reading everybody. Vizard Agent casts per character, fits each new line into its existing window, and rebuilds the captions, because every change to the dialogue invalidates the timings they were built from. In the session behind this article it:
- Found two distinct voices for the two male characters.
- Generated each character's dialogue separately and refined each line so it fitted its scene naturally.
- Generated replacements aligned to the existing timeline windows and rebuilt the mixed track using those windows, so nothing downstream drifted.
- Transcribed the mixed track rather than reusing the old transcript, and built the captions from that transcription.
- Applied strict gender assignment across the whole dialogue track.
- Inspected any shot where it was unclear who was speaking, rather than guessing.
When does this not work well?
Some dialogue simply cannot be separated. Overlapping speech, crowd scenes and any passage where two people share one track will not divide cleanly into per-character lines, and Vizard Agent will say so rather than force it. Characters you never see and strict lip sync are hard too.
Common questions
Can two men sound properly different? Yes, if you ask. Vizard Agent chooses for contrast rather than just gender.
Will the lines still fit? Yes. Vizard Agent generates each into its existing timing window.
Do the captions change? They must. Vizard Agent rebuilds them from the new dialogue track.