How to transcribe a video into a dialogue script with the speaker names
Ask Vizard Agent for the dialogue laid out as a script rather than as a transcript. It extracts the spoken words with timings, pulls detailed frames of each character to see who is moving and speaking at each line, matches every sentence to a speaker, and writes the result with names against the lines, translated or in the original, as you prefer.
What do you type into Vizard Agent?
Ask for a script and state the language rule in the same instruction. A transcript and a script are different deliverables, and a video with two languages in it will otherwise come back with both tangled together on the page. If you know the characters' names, give them; otherwise Vizard Agent describes them.
Prompt
What does Vizard Agent actually do?
Transcription gives you the words and nothing else. Who said each of them turns out to be a question about the picture rather than the audio. So the transcription is quick, and almost all the effort goes into reading mouth movement, gesture and who is facing the camera in fine frame strips. In the session behind this article it:
- Sampled frames to see what was in a two-character animated clip, then extracted the spoken words with timings before any attribution started.
- Reviewed the characters and pulled a detailed preview of their movement.
- Inspected those frames to see who was speaking on each line.
- Matched the speaker across the first half and then the second half separately, so one wrong attribution could not carry on to the end.
- Wrote the dialogue file with names against each line, then produced a single-language version.
When does this not work well?
Attribution needs something visible to go on. Off-screen speakers, crowd scenes, overlapping dialogue and heavy stylisation all make it hard to tell who is talking, and a narrator over action is nobody on screen. Vizard Agent will say when it has nothing, because a guess in a script is worse than an honest gap.
Common questions
Why not use the audio to tell speakers apart? It helps, but Vizard Agent uses the picture to settle the cases audio cannot.
Does it work on animation? Yes, and there the picture is often the only reliable signal.
What if two people talk at once? Vizard Agent marks it rather than guessing at a split.