How to caption the sounds and the speakers, not just the words
Ask Vizard Agent for speaker-labelled captions with sound cues, not just a transcript on screen. It works out who is speaking by sampling the mouths in the picture at each line rather than trusting the audio separation alone, and adds cues for the sounds that carry meaning, such as a knock or a car pulling away, while leaving the ambient ones out.
What do you type into Vizard Agent?
Ask Vizard Agent for the two additions by name. A phrase like "captions for the deaf" means slightly different things in different places, and spelling out "speaker labels and sound cues" removes that ambiguity in one sentence. Give the names of the people too: it can tell two voices apart, but only someone who has watched the film knows which one is Maria.
Prompt
What does Vizard Agent actually do?
Audio separation tells you there are two voices. Looking at whose mouth is moving at that moment tells you which one is which, and getting that wrong is worse than having no labels at all. Sound cues don't appear in a transcript, so they come from a separate listening pass. In the session behind this article it:
- Listened through a short film for dialogue, music and sound-effect events before building anything.
- Built timed captions from the dialogue transcript and inspected its structure for speaker information.
- Sampled the visible mouth movements at each dialogue cue and reviewed the speaker frames across every section of the video.
- Built speaker-labelled dialogue and sound-cue caption data and tested it on the opening section.
- Checked the labels and cues for readability at real size, since both take room on the line, then moved the caption block into the visible safe area and rendered the full video.
When does this not work well?
Attribution needs faces to check against. An off-screen voice, a crowd or a conversation shot from behind gives Vizard Agent nothing to verify, so those lines are labelled from context or left unlabelled rather than guessed. Overlapping speech, dense sound design and very fast dialogue are hard too, because labels and cues crowd the line.
Common questions
Which sounds should be cued? The ones that carry meaning. Vizard Agent leaves ambience out unless you ask.
What if the speaker is off screen? Vizard Agent flags it rather than guessing, and you decide how to label it.
Can it produce a sidecar file? Yes. Vizard Agent delivers one as well as, or instead of, burnt-in captions.