Vizard Agent

How to stop an edit feeling like a slideshow

Last updated 2026-09-05 · 8 min read

Tell Vizard Agent the cut feels like a slideshow and point at a moment where it pauses. It rebuilds those shots from the moving footage, then measures the real motion in every shot and samples each one's first and last frame — because a frozen shot and a very still one look identical until you compare their endpoints.

What is the short version?

A video made of held frames is a slideshow with a soundtrack, and the giveaway is a face that stops moving mid-sentence. The fix is to use the footage of people talking rather than a frame of them talking, and then to check that every shot actually moves.

  1. Go to Vizard Agent with the cut.
  2. Say it feels like a slideshow and where it pauses.
  3. Ask for the moving footage of those people instead.

What do you need before you start?

The cut, and one moment you can point at. A timestamp is ideal but a description works — "it pauses on the first woman's face while she is talking" locates the problem precisely enough, and the exact shot gets found from there.

What do you type into Vizard Agent?

Describe the feeling and then give an example of it. "More dynamic" is not actionable, but "it pauses on people's faces while they are talking, which looks odd" is a specific fault with a specific cause and a specific repair.

Prompt

This pauses on people's faces while they are talking, which looks odd — it feels like a slideshow rather than a video. Can you use the footage of them speaking instead of stills of them speaking? There is also just a picture of empty chairs around 35 seconds.

Variants worth knowing:

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real forty-five-second event reel that had come back feeling static. The last third is a verification method rather than an edit, and it is the part worth carrying away.

  1. Counts frames per shot to see how long each one holds.
  2. Rebuilds the cut with a faster, intercut structure.
  3. Re-times the captions to the new shot lengths.
  4. Reviews every shot of the new version individually.
  5. Judges the two weakest slots and replaces those shots.
  6. Finds the frozen-frame fault in how the shots were being built.
  7. Re-renders with live motion throughout.
  8. Measures the real motion per shot in the new render.
  9. Samples the start and the end frame of every shot.
  10. Reviews the live version shot by shot to confirm.
  11. Finds a fuller framing for a shot that still felt empty.
  12. Renders and checks the finished frames one last time.

Steps eight and nine are the method. Comparing a shot's first frame against its last frame is a direct test of whether anything happened in it, and it catches a frozen shot that a person watching at speed will simply register as "a bit slow".

Step three is easy to lose track of. Every time the shot lengths change the captions have to move with them, so Vizard Agent re-times them against the rebuilt cut rather than carrying over timings that belonged to the version before it.

Step five is the other half. Some shots are not frozen, they are just empty — a row of chairs, a wall, a wide with nobody in it — and no amount of motion rescues them. Those get replaced rather than fixed.

What does the result look like?

A cut where people move while they speak, the wide shots have something happening in them, and no frame is held long enough to read as a photograph. The captions still line up because they were re-timed to the shots that survived.

It is the same footage and the same length. It simply plays like video.

When does this not work well?

You need moving footage to cut to. Where the only coverage of somebody is a photograph, or where the moving version is unusable — out of focus, badly framed, someone walking through — a still is the honest choice and the fix is to hold it deliberately rather than accidentally.

How do you fix a result that came back wrong?

Point at the shot and say what it does. Vizard Agent keeps the per-shot motion measurements and the endpoint frames, so "this one still sits there" resolves to a specific shot rather than a general impression of the pacing.

How does Vizard Agent compare to doing it yourself?

By hand this is judged by watching, and watching is exactly what fails here — a frozen shot inside a music-led edit reads as a pause in the pacing rather than as a fault. The endpoint comparison takes seconds and is unambiguous.

By hand Vizard Agent
Spotting a frozen shot Noticed as "slow" First and last frame compared
Judging the pacing By feel Frames counted per shot
Empty shots Left in Judged and replaced
The repair Trim it shorter Rebuilt from moving footage
Captions afterwards Drift Re-timed to the new lengths

Common questions

Why did it use stills at all? Sometimes coverage is thin. Vizard Agent will say which shots had none.

Can it check every shot moves? Yes. Vizard Agent measures motion in each one.

Will the video get longer? No. Vizard Agent swaps shots rather than adding them.

What about the empty wide shots? Say which. Vizard Agent replaces them with better coverage.

Can I keep one or two stills? Yes. Name them and Vizard Agent holds those deliberately.

Will the captions still fit? Yes. Vizard Agent re-times them to the final cut.

Can it intercut faster instead? Yes, and Vizard Agent will if the shots are long.

What if the moving footage is shaky? Vizard Agent can stabilise it or pick a different take.

Does this work on a photo montage? Differently. There the fix is movement within each photo.

Can it show me which shots were frozen? Yes. Ask Vizard Agent for the per-shot measurements.

What about a slow-motion shot? That still moves. Vizard Agent measures motion, not speed.

Will the music need re-timing? Only if the shot lengths change, and Vizard Agent handles it.

Can it do this on a delivered file? Better from the project, but Vizard Agent can diagnose the export.

Why measure rather than watch? Because a frozen shot inside a paced edit feels like a slow moment, and a slow moment is something people talk themselves out of noticing.