How to have the text on screen read aloud in your video
Upload the video and ask for the on-screen text to be narrated. Vizard Agent reads the captions off the frames themselves, detects the exact moment each one appears and disappears, records a line for each, and delays every one to land with its caption.
What do you type into Vizard Agent?
Say what to read and how it should sound. Vizard Agent works out the timing itself by detecting when each caption appears, so the instruction that matters is about the voice and about what happens to the audio already in the video: kept underneath at a lower level, or dropped.
Prompt
What does Vizard Agent actually do?
A video with burnt-in captions has no subtitle file and no transcript, so there is nothing to hand over except the picture. Vizard Agent establishes both the text and its timing from the pixels, and getting the crop band right is what turns a smear of half-captions into a clean list of lines with times. In the session behind this article it:
- Probed the 36.43-second, 1080x1920 video, tiled frames into a grid to find the text, and confirmed there was no existing script or subtitle file.
- Wrote a script to detect the exact timestamps at which the captions changed, cropping the text area every half second.
- Located the text's bounding box, narrowed the crop to the band where it sat, and verified every caption's wording and its appearance and disappearance times.
- Searched the voice library for a professional narrator, generated the eleven narration segments in parallel and measured each one's duration.
- Measured the original audio's loudness, then delayed each segment to its caption's start and mixed it over the original audio.
- Verified the render's sync, narration clarity and background balance.
When does this not work well?
Reading text out of pixels is harder than reading it out of a file. Text over busy footage, decorative or animated type, very fast captions and stylised fonts all make it harder still. Vizard Agent tells you which captions it could not resolve rather than guessing at a word and narrating something you never wrote.
Common questions
Do I need a subtitle file? No, and that is the point. Vizard Agent reads the captions off the picture itself.
What if the video has audio already? It stays. Vizard Agent mixes the narration over it at a measured level.
Will the lines overlap? Not if the timings allow. Vizard Agent measures each segment before placing it.