How to change specific text in a video while keeping the animation
Upload the video to Vizard Agent and say which word should become which. Vizard Agent zooms into the text, draws a coordinate grid over the frame to measure exactly where the letters sit, renders a comparison of the available fonts to match the original type, and replaces the word without disturbing the animation around it.
What do you type into Vizard Agent?
Quote both words exactly as they should read, capitalisation included, because the new one goes in as written. Vizard Agent locates the text visually rather than needing coordinates from you, and the precise replacement wording is what lets it match the letter spacing and size rather than approximate them. If the word appears more than once, say whether every instance changes.
Prompt
What does Vizard Agent actually do?
A word rendered into the picture isn't text any more. It's pixels that happen to look like letters. So most of the job is measurement: Vizard Agent doesn't eyeball the position, it draws a grid and reads the numbers off it, and it replaces only the region the text occupies. In the session behind this article it:
- Pulled a full-resolution frame, zoomed into the screen the text sat on, and probed and downloaded the source.
- Analysed the frame-to-frame movement and checked for scene changes, to see whether anything shifted under the text.
- Drew a coordinate grid over the text area to read off its exact position, and a second grid over the whole object to place it in context.
- Searched the installed fonts and rendered a comparison to match the original type, then measured the letter size and bounding box.
- Tested the replacement on one frame and compared four different replacement methods before committing to one.
- Delivered a 576x1024, 30fps file of 24.05 seconds, the same size and length as the original, with only the text region changed.
When does this not work well?
Text baked into a picture has no editable layer, so every replacement is a reconstruction. If the font isn't installed, the closest match can be visibly different, a much longer word changes the spacing around it, and text moving with a handheld camera is far harder than text on a static screen. If you still have the original project file, changing the word there is always cleaner.
Common questions
Does it work if the text is part of the picture? Yes, that is the case this is for. Vizard Agent reconstructs the region rather than editing a text layer.
Can it match the original font? Vizard Agent renders a comparison of the installed fonts and picks the closest. An unusual typeface may not have a match.
What if the new word is longer? Say so. Vizard Agent rebuilds the spacing, and a much longer word changes the layout around it.