Vizard Agent

How to cut a wedding reel to the band's own performance

Last updated 2026-08-30 · 7 min read

Upload the clips and tell Vizard Agent what the night should feel like. Vizard Agent watches every clip for the song, the energy and the best crowd moments, works out which ones share a take, detects the beat grid, and cuts so the edit lands on the lyric rather than beside it.

What is the short version?

A wedding reel that cuts on the music feels like being there. One that cuts on a timer feels like a slideshow of a party. The difference is knowing where the beats and the lyrics actually fall in the footage you have.

  1. Go to Vizard Agent and upload every clip from the night.
  2. Say what it should feel like rather than what to include.
  3. Say the length — twenty-five to thirty-five seconds.

What do you need before you start?

The clips and a feeling to aim at. Vizard Agent identifies the songs and rates the energy in each clip itself, so no shot list is needed from you. Describing the atmosphere you want does considerably more work than listing the moments you happen to remember.

What do you type into Vizard Agent?

Describe the atmosphere rather than the shots. Vizard Agent selects for energy and crowd engagement when you describe a feeling, which produces a very different and better reel than a list of the moments you happen to remember from the night.

Prompt

Create one polished [25-35] second vertical hero reel from this wedding footage. The goal is to show what the night actually felt like: a packed dance floor, guests fully engaged, and the band driving the party.

Variants worth knowing:

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real wedding hero reel built from eleven clips. The step that makes the cutting work is the ninth: it checked whether three clips of the same song were actually the same take, because that decides whether they can be intercut.

  1. Takes a first look at all the clips — length, size and orientation.
  2. Downloads them and pulls sample frames from each, one by one when the batch fails.
  3. Combines the sample frames into one contact sheet and reviews it.
  4. Checks how the analysis tools work before using them.
  5. Watches each clip to identify the song, the energy level and the best crowd moments.
  6. Watches the remaining clips the same way.
  7. Checks whether the clips of the same song come from one take, and finds where they overlap.
  8. Gets word-level timing of the performance so the cuts can land on the lyric.
  9. Pulls frames from each candidate window, detects the beat grid, regenerates a missing strip, reviews all the candidates side by side, splits the review into readable stacks and fixes a clip whose frames would not extract.

Step 7 is the one that decides the shape of the reel. Three angles of the same take can be intercut into one continuous performance; three separate performances of the same song cannot, and mistaking one for the other produces a reel where the music jumps.

What does the result look like?

The clips this was built from measure 1920x1080, HEVC, 30fps, the longest running 64.9 seconds. From eleven of them comes a vertical hero reel of under half a minute: the floor filling, the band driving, the crowd on the beat, and the cuts landing on words rather than between them.

Cutting to the lyric is what sells it. A reel cut on the beat alone is rhythmic; a reel that changes shot on the word everybody in the room was singing feels like the night rather than a recording of it.

When does this not work well?

A reel about atmosphere needs that atmosphere to have been captured on the night, and phone footage of a dark, loud room has real limits that Vizard Agent cannot edit away. These are worth knowing before you promise anyone a hero reel.

How do you fix a result that came back wrong?

Name the moment or the song that is wrong. Vizard Agent keeps the per-clip analysis, the beat grid, all the word timings and the candidate review stacks, so the selection can change without re-watching any of the footage.

How does Vizard Agent compare to doing it yourself?

By hand this means watching eleven clips of a loud room, working out which are the same song, tapping out a beat grid, and then discovering that the good crowd shot is from a different take and will not cut against the rest.

By hand Vizard Agent
Identifying songs Listen through everything Identified per clip, with energy rated
Same take or not Guess from the music Checked and overlaps found
Cut points Tap along Word-level timing plus a beat grid
Choosing shots Scrub and settle Candidates reviewed side by side

Common questions

How many clips do I need? Eleven made a strong reel here. Fewer works if the energy is in them.

Does Vizard Agent know which song is playing? Vizard Agent identifies the song in each clip and rates the energy, which is how it groups them.

Why does the same take matter? Because only clips from one continuous performance can be intercut without the music jumping.

Can it cut to the words? Yes. Vizard Agent pulls word-level timing so the shot changes land on the lyric itself.

What about the dark and noisy footage? It works with what it has. Vizard Agent selects the cleanest moments rather than fixing the room.

Should the reel be about the couple or the band? Say which. A reel for the band leads on the crowd; a reel for the couple leads on them.

Can I get a longer version? Yes. Vizard Agent builds one from the same analysis and the same beat grid.

Does it use the phone audio? Where it holds up. Vizard Agent checks it rather than assuming a loud room recorded well.

What about guests who did not consent? That is your call. A dance floor contains people who did not agree to be posted.

Can one night produce several reels? Yes, and it is the efficient way to work. Vizard Agent already holds the per-clip analysis and the beat grid, so a second reel for a different song costs a fraction of the first.

Does Vizard Agent check the finished reel? Yes. It reviews the candidates and the cut alignment before delivery.