How to time everyone's attempt in a group challenge video

Last updated 2026-09-01 · 8 min read

What is the short version?

The fun of a group challenge is the comparison, and the comparison only works if everyone was timed the same way. Eyeballing it puts half a second on somebody's total, and half a second is usually the whole result.

1. Go to Vizard Agent and upload each person's clip.
2. Say exactly what starts the clock — first touch, first step, the whistle.
3. Say exactly what stops it — the lockout, the line, the release.

What do you need before you start?

Everyone's attempt, and a definition of the start and finish that all of you would still agree on afterwards. Vizard Agent measures strictly against whatever you specify, which makes the definition the entire fairness question — first touch and first movement are genuinely different moments.

• Every clip. One per attempt, uncut.
• The start. The precise action, not "when they begin".
• The finish. The precise position that ends it.
• The names. For the on-screen times.
• The order. Chronological, or ranked by time.

What do you type into Vizard Agent?

Define the two moments as precisely as you can bear to. Vizard Agent goes and finds them frame by frame in each clip, so the difference between a comparison everyone accepts and an argument at the pub afterwards lies entirely in how tightly you described the start and the finish.

Prompt

Variants worth knowing:

• Define both ends. Vague ends produce unfair times.
• Say "the same for everyone." It is the point of the video.
• Name the people. So the times are labelled.

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real five-person challenge. Almost the entire session goes on locating just two frames per person, which sounds wildly excessive right up until you realise that those ten frames between them are the whole result everybody is going to argue about.

1. Checks the workspace for the clips and assets available.
2. Probes each clip and extracts thumbnails to see its contents and duration.
3. Builds a visual grid of all the thumbnails and reviews it to identify the people and the movement.
4. Rebuilds the grid at the correct proportions and reviews the timings of each participant.
5. Extracts frame-by-frame images at 0.1-second intervals across all five clips.
6. Generates high-precision grids at 0.3 and 0.5-second intervals with the timestamps drawn onto them.
7. Analyses clip one at half-second intervals to find its start and finish.
8. Analyses clips two to five at three-tenths of a second for the same two moments.
9. Extracts frames around each attempt's start and end at the finest interval.
10. Builds labelled contact sheets for the start and end of every successful attempt.
11. Reviews each clip's start frame to find the exact moment of first contact.
12. Reviews each clip's end frame to find the exact finishing position.

Step six is the trick worth stealing. The timestamps are drawn onto the contact sheet itself, so reading a time off a frame is a matter of looking rather than of counting frames and doing arithmetic.

What does the result look like?

The challenge this page is written from was five clips, each analysed down to tenths of a second, with the start and finish of every attempt located on a specific frame and verified on a labelled contact sheet before anything was cut.

The point of all that is a number on screen you can defend. When the gap between first and second is a fraction of a second, an estimated time is not a result, it is an opinion.

When does this not work well?

Frame-accurate timing needs the deciding moment to be genuinely visible in the footage, and phone video shot by a group of friends messing about is almost never filmed with that requirement in mind. These are the cases where it cannot give you a defensible number.

• The start is off camera. Nothing recovers it.
• Low frame rates. Thirty frames a second limits precision.
• Motion blur. The exact contact frame is ambiguous.
• Different camera angles. They can disagree about the moment.
• Vague definitions. They make the comparison unfair, not the tool.

How do you fix a result that came back wrong?

Say which attempt you are disputing and which end of it. Vizard Agent keeps all the frame grids and the labelled contact sheets it worked from, so a re-judged start or finish is a lookup in material it already has rather than another pass through the footage.

• "He started earlier than that." Re-judged from the same grid.
• "That is not full lockout." Finish redefined and all clips re-measured.
• "Rank them by time." Re-ordered from the measured results.

How does Vizard Agent compare to doing it yourself?

By hand this means scrubbing frame by frame in an editor for ten separate moments, and in practice it means somebody guessing and everyone else disputing it. Vizard Agent measures each moment and shows you the frame it used.

Common questions

How precise can it get? To the frame. Vizard Agent steps through at a tenth of a second and narrows from there.

What if the frame rate is low? That is the limit. Vizard Agent cannot resolve finer than the footage allows.

Can it use different definitions per person? It can, but it should not. Vizard Agent applies one definition so the comparison holds.

What if someone failed the attempt? Say so. Vizard Agent measures only the attempts you count as successful.

Can it put the times on screen? Yes, labelled with names. Vizard Agent places them against each clip.

Can it rank them? Yes. Vizard Agent orders the clips by the measured times if you want a leaderboard.

Does everyone need the same camera? No, but it helps. Vizard Agent notes when angles disagree about the moment.

What if the start is blurred? It flags the ambiguity. Vizard Agent shows you the candidate frames rather than picking silently.

Can I check its judgement? Yes, and you should. Vizard Agent keeps the labelled contact sheets it read the times off.

Does it work for a race? Yes. Vizard Agent handles any challenge with a defined start and finish.

Why does the definition matter so much? Because "when they start" and "when they first touch it" can be half a second apart, and half a second is usually the difference between first and third. The tool cannot make that choice for you.

Can it add music and captions? Yes. Vizard Agent cuts the attempts together with the times as on-screen graphics.

Does it check its own measurements? Yes. Vizard Agent re-extracts frames around each located moment and reviews them before using the time.