How to make two labels on a slide change at the same moment

Last updated 2026-09-15 · 8 min read

What is the short version?

A slide that shows a word and then its changed form is doing two jobs at once, and the animation usually swallows the first one. The viewer sees motion, then the second form, and never actually read the first.

1. Give Vizard Agent the slide and say which two states matter.
2. Ask it to hold the first state, then change both labels together.
3. Ask for each state as a still to check.

What do you need before you start?

The video with the slide in it and a clear statement of what the two states are. Vizard Agent can see the animation, but only you know that the point of this slide is a change from one form to another rather than a decorative transition, and saying so changes what it looks for.

• The video. The finished one, with its animation.
• The two states. Written out, exactly.
• Which labels are paired. They change together.
• Where each one sits. Top line, bottom line.
• The original look. It must survive the fix.

What do you type into Vizard Agent?

Ask for states, not speed. "Slow it down" is the obvious request and it is usually the wrong one, because the first state may not exist as a clean frame anywhere in the source — slowing the animation just gives you a longer look at the smear rather than a readable word.

Prompt

Variants worth knowing:

• "Two clear held states." What you actually want.
• "Both labels on the same frame." Stops a stagger.
• "Keep the original animation." Elsewhere, untouched.

What does Vizard Agent actually do?

Here is the sequence on a real grammar slide whose opening form was unreadable. Notice how much of it is spent looking for a clean frame before any rendering happens — that search is the actual work, and skipping it is why the obvious fixes fail.

1. Extends the readable first form, then slows the original animation.
2. Checks whether the slowed version is now fully readable.
3. Searches the source for the exact readable opening state.
4. Reads the original frames of that scene for a clean state.
5. Renders a version built on the genuinely readable frame.
6. Compares every early animation state to pick the right one.
7. Builds a corrected opening frame when none of them is clean.
8. Measures the position, size and style of the label to replace.
9. Replaces only the wrong label, in the matching style.
10. Verifies both visible forms in the rendered result.

Step six is the step people skip. An animation can pass through a dozen intermediate states, several of which look clean in motion and are not, so Vizard Agent compares them as stills rather than trusting any single one.

Step seven is the fallback that makes the job finishable. If the source genuinely never holds the first form, Vizard Agent constructs that frame rather than reporting that it cannot be done, and the rest follows from there.

Step nine is why the repair is invisible. Vizard Agent measures the existing label first, so the replacement sits at the same position in the same style and the slide does not suddenly look like it was edited.

What does the result look like?

Two states a viewer can read, with the same animation around them. The first form holds long enough to land, both labels change on one frame, and the slide still moves the way the original did. Vizard Agent reports the timing of each state so you can check it without scrubbing.

When does this not work well?

Some slides are built so that the change is the animation. If the two forms only ever exist as endpoints of a continuous morph, holding one means creating a frame that was never in the video, and Vizard Agent will say so rather than pretend it found one.

• Continuous morphs. No clean state exists to hold.
• Very short slides. Holding the first form eats the second.
• Heavily stylised type. The rebuilt label may not match.
• Labels that are part of an image. Not text, harder to replace.
• Narration locked to the old timing. Holding pushes it out of sync.

How do you fix a result that came back wrong?

Point at the state rather than the moment. Because Vizard Agent now treats this slide as two named states instead of one transition, saying "the first state still is not clean" is something it can act on directly, and a fix to one state does not disturb the other.

• "The first form is still smeared." A cleaner frame is built.
• "They change one after the other." Both are put on one frame.
• "The new label looks different." Position and style are remeasured.
• "It holds too long now." The hold is shortened, the change kept.

How does Vizard Agent compare to doing it yourself?

By hand you slow the clip down, watch it, and think it is fixed, because at half speed your eye fills in the word it expects. The problem only shows up when somebody who does not already know the answer watches it and asks what the first word was.

Common questions

Why is slowing it down not enough? Because a slow smear is still a smear. Vizard Agent looks for a clean frame instead.

How does it find the readable frame? Vizard Agent compares the early animation states as stills rather than in motion.

What if no frame is clean? Vizard Agent builds one from the nearest state and the correct text.

Will the rest of the animation survive? Yes. Vizard Agent changes the one scene and leaves the others as they are.

Can both labels really change on one frame? Yes, and asking for it explicitly is what stops Vizard Agent staggering them.

What if the labels are different sizes? Vizard Agent measures each one separately, so both keep their own style.

Does this work for numbers? Yes. A figure changing to another figure is the same problem.

Can it handle three states? Yes, though tell Vizard Agent the order, or it will infer one.

Will the video get longer? Slightly, because the held state takes time. Vizard Agent reports the new duration.

What about the narration? Say if it must stay in sync, and Vizard Agent paces the hold to fit it.

Can it check the fix for me? Yes. Vizard Agent verifies both forms in the rendered file, not the timeline.

What if the wrong form was there originally? Vizard Agent replaces just that label in the matching style.

Does the slide look edited afterwards? It should not, which is why the style measurement in step eight matters.

Should I describe the states or the timing? The states. Vizard Agent works out the timing from them.