Shoot a Small Stop-Motion Test for a Commercial Treatment

A stop-motion test is not your storyboard panels held on screen for different lengths of time. It photographs new states of a real object and plays them back at a rate you chose. That difference is the reason it earns its place in a treatment: a timed storyboard shows you when things change, while a stop-motion test shows you what a change looks like when it's built out of discrete positions.

Which also means you can't judge it from the stills. You have to watch it run.

The recipe is short. Lock the camera, choose an action small enough to finish, move the object a deliberate distance between exposures, then look at the sequence at the rate you recorded. One motion question per test. Not a rehearsal of the production plan, and not a budget.

Choose an action small enough to test completely

An action you can finish beats an action that reads well on paper. For a treatment, that means a start arrangement, an end arrangement, and one change between them: advancing, hesitating, turning, coming to rest. If you can't say what the last frame looks like, you aren't testing an action, you're filling a card.

Take something off the table in front of you. A mug works well, because the handle makes rotation legible — you can see a turn in the handle's position instead of inferring it from a shadow.

Set it up like this. Lay a strip of low-tack tape on the tabletop and mark three points: a start mark, an end mark 12 cm to the right, and a midpoint at 6 cm. Put the mug on the start mark with the handle pointing to camera right. The action is: the mug advances to the end mark, pauses, then turns 90° clockwise so the handle points away from camera.

That gives the test everything it needs. Two arrangements, one change, a distance you can check with a ruler, and a rotation you can check by eye.

Leave fabrication and rigging out of it unless the rig is genuinely the thing you're deciding about. A puppet with armature wire and replacement mouths is a different project with a different first test, and the failure modes of the rig will drown out the failure modes of your spacing. The mug doesn't need to be interesting. It needs to be complete.

Stabilize the frame before recording the sequence

Camera on a tripod. Manual focus, manual exposure, manual white balance. Not because manual is virtuous, but because anything automatic will change between exposures, and change is exactly what you're trying to measure.

This matters more than it sounds, because camera movement and object movement look the same on playback. If the tripod settles half a centimetre mid-sequence, the whole scene slides and your eye reads it as the mug doing something strange — usually a hitch or an overshoot. You'll then go hunting for the problem in your spacing, where it isn't.

So before you shoot anything you plan to keep, take two exposures with nothing moving. Don't touch the table, don't touch the camera, don't switch the lamp. Compare them. If they differ, something is drifting: the tripod, daylight moving across the room, or auto exposure hunting as the mug's shadow crosses the metering zone. Find it before you shoot twenty-four frames and blame them on your spacing.

Lock the light as well. A window is a moving light source, and mixed window-plus-lamp setups shift colour as the afternoon goes. If the window has to be part of the setup, write down the time you shot.

Frame wide enough to hold both ends of the action. The mug needs to sit inside the frame on the start mark and on the end mark, with the whole 90° turn visible and the handle readable at both. Then check focus at both marks. If the mug is soft at the end mark, stop down a little or pull the camera back.

Write the setup down before you shoot, not afterwards: camera, lens, aperture, shutter, ISO, playback rate, distance to subject, lighting. Those numbers are the only thing that makes two versions comparable.

Use previous frames to control the next increment

The Stop Motion Studio help page, checked on 2026-09-09, documents overlay opacity and prior-frame layers that compare the live view with earlier frames during capture. Dragonframe's feature page documents frame-based timing and reshoot operations. Both are vendor descriptions of functions rather than a workflow anyone in this article has run, and naming them is not a recommendation to buy anything — but the overlay function is the part you'll lean on here.

What the overlay does: it puts one or more earlier frames over the live view at reduced opacity, so you can see where the object sits now against where it sat at the previous exposure. That's a positioning aid. You use it to place the next position on purpose instead of by feel.

What it doesn't do: tell you whether the spacing is any good. The overlay is a still comparison, and you are making a moving thing. Positioning and motion judgement are two different jobs done at two different times. Do the first at the table. Do the second at playback.

Here's the distinction that trips people up. Two things both change how fast the action seems to go, and they are not the same thing:

• Spacing — how far the object moves between exposures.
• Duration — how many exposures a position is displayed for, and how many exposures play per second.

If each exposure plays once, apparent speed is step size multiplied by playback rate. A 1 cm step at 12 exposures per second reads as 12 cm per second. A 0.5 cm step at 24 per second reads as the same 12 cm per second. Same average, different grain: twelve larger displacements versus twenty-four smaller ones. The arithmetic is certain. What the difference feels like is precisely the kind of question a test settles and an argument doesn't.

The test, written down

Two routes, one change between them. The mug is placed on the start mark before either route, and both end with it 12 cm right and turned 90°. Both are 24 exposures. Both play at 12 exposures per second with one exposure per played frame — shooting "on ones" — so both run two seconds.

Read each row below as the displacement you apply before you take that exposure, not as the position the exposure records. The start mark is where you set the mug down, not a frame in the film: exposure 1 is already 1.0 cm in on Route A and 1.5 cm in on Route B. The end mark is recorded — exposure 24 sits on it — and the 6 cm midpoint gives you a cross-check partway through, recorded at exposure 6 on Route A and exposure 4 on Route B.

Route A moves the mug the same distance on every exposure. Route B covers 9 cm in its first six exposures and 3 cm in its last six — 18 cm/s, then 6 cm/s — so the arrival slows into the pause. Nothing else differs. Same distance, same exposure count, same travel time, same average speed across the travel. The hold is six exposures in both routes, half a second of a mug that isn't moving. The turn is identical.

That's the reason to write it down this way. If Route B reads differently on playback, you can't blame the distance, the duration, the hold, or the turn. All that's left is where the small steps are.

These are specifications, not a shoot report. The frames don't exist yet; no one has captured them. That isn't a dodge — it's the value of giving you the numbers. You can check the arithmetic with a ruler and a stopwatch, shoot both routes in an afternoon, and find out whether the prediction holds. The prediction, stated plainly, is that a pause reads as a decision when the approach decelerates into it and as an interruption when the approach arrives at full speed. It may be wrong. It has to be able to be wrong, or it isn't a test.

Play the complete action before calling it successful

Play at the rate you wrote down. Twenty-four exposures at 12 per second is two seconds. Watch it three or four times before you diagnose anything.

Things worth watching for, in roughly the order they show up:

• Does the pause register at all, or does it read as a dropped frame?
• Does the mug look like it stops, or does the footage look like it stops?
• Any hitch where the camera drifted
• Any brightness change, which is usually exposure or light rather than spacing
• Does the turn start cleanly out of the hold, or do the pause and the turn fight each other?
• Does the mug stay inside the frame for the whole travel?

Now the part people get wrong. A contact sheet of your 24 exposures is a useful object. It shows you positions. It will show you immediately that Route B's first six steps are visibly wider than Route A's. That's a real check and worth making.

It cannot show you the motion, because the motion isn't in the pictures. It's in the relationship between the pictures and the rate. Twenty-four correctly spaced stills played at the wrong rate is a different piece of film. Judging the test off thumbnails tells you about your spacing and nothing about your timing, which is roughly half of what you asked.

A note on the hold itself, since it appears in both routes. If you re-photograph the same position six times, the mug is in the same place in all six — unless someone leaned on the table, in which case the drift tells you something useful and unwelcome. If instead you expose once and repeat the frame in the timeline, the six frames are exactly identical and no drift is possible, but the decision has moved into post. Pick one route, note which you used, and keep it the same across both versions.

Revise spacing while preserving the record

Change one thing. Shoot Route B after Route A. Keep Route A's source frames and its sequence order. Don't overwrite, don't delete, don't tidy the folder.

If you delete Route A, what you have left is a version that looks better than nothing, and better than nothing isn't evidence. You also lose the ability to show anyone what the choice was. An argument in a treatment is stronger when you can put two versions side by side and name the change between them, because the change is the claim.

The duplication trap is worth naming. A deliberate hold and an accidental duplicate are identical in the timeline. Deliberate: you meant the mug to sit still for six frames. Accidental: you pressed the shutter twice, or you nudged the mug and put it back, or you decided the exposure hadn't fired and took it again. All three produce a repeated position. Write your holds down as you shoot, in the order they happen, or three weeks later you'll be guessing whether the stutter at frame 14 was a decision.

Record the settings alongside each version: playback rate, aperture, shutter, ISO, lighting, clock time. If the lamp changed or the sun moved, you can't compare the two, because flicker looks exactly like a spacing problem and isn't one.

One confound deserves naming out loud. You shoot the second version with a steadier hand than the first, so it will look better for that reason alone. A cheap defence is to shoot A, then B, then A again. If the two A passes match each other, you have some grounds for saying B is genuinely different. If they don't match, you've learned that your setup isn't repeatable yet — which is worth knowing before you promise anyone anything.

What the test carries, and what it doesn't

What you end up with, when you shoot it, is two two-second sequences that differ in one respect you can point at. That is a modest thing and a real one. In a treatment it lets you say something specific about how a movement reads and have it be true, and it lets you play the evidence for someone rather than describing it.

What it doesn't give you is anything about labour, cost, crew, or schedule. A mug on a tabletop is not a production plan. It says nothing about how a rig will hold over four hundred exposures, whether a puppet survives a shoot day, how many replacement parts you'd need, or how long any of it takes. Vendor feature pages describe functions; they don't forecast your room, your tripod, or your afternoon.

And if the two routes read the same, that's a result too. Route A and Route B differ in exactly one place — where the small steps sit. If the pause lands identically in both, then decelerating into a hold doesn't buy you what you thought it bought, and you've saved yourself from arguing about it in front of anyone. Say that in the treatment. A test that comes back negative is still the reason you know.

Frequently asked questions

How is a stop-motion test different from a timed storyboard?

A stop-motion test photographs new states of a real object and plays them back at a chosen rate. A timed storyboard shows when things change; the test shows what a change looks like when built out of discrete positions. You cannot judge it from stills, because the motion is in the relationship between pictures and rate.

What action should a small stop-motion test use?

Choose an action you can finish completely: a start arrangement, an end arrangement, and one change between them, such as advancing, hesitating, turning, or coming to rest. The fixture uses a mug, tape marks at 0 cm, 6 cm, and 12 cm, a 12 cm advance, a pause, and a 90-degree clockwise turn. Leave fabrication and rigging out unless the rig is the thing you are deciding about.

Why test camera stability and lighting before shooting the sequence?

Camera on tripod, manual focus, manual exposure, and manual white balance, because automatic changes between exposures are what you are trying to measure. Camera movement and object movement look the same on playback, so a settling tripod can read as a hitch or overshoot and send you hunting in your spacing. Take two exposures with nothing moving and compare them; if they differ, find the drift before shooting many frames. Lock the light too, and note the clock time if a window is part of the setup.

What does an overlay or prior-frame layer do, and what does it not do?

It puts one or more earlier frames over the live view at reduced opacity so you can see where the object sits now against its previous position. That is a positioning aid. It does not tell you whether the spacing is good. Positioning and motion judgement are different jobs at different times: do the first at the table and the second at playback.

How do spacing and duration affect apparent speed, and what does the Route A and Route B test compare?

Spacing is how far the object moves between exposures. Duration is how many exposures a position is displayed for and how many exposures play per second. If each exposure plays once, apparent speed is step size multiplied by playback rate. A 1 cm step at 12 exposures per second reads as 12 cm per second; a 0.5 cm step at 24 per second reads as the same average speed with a different grain. The Route A and Route B test uses 24 exposures at 12 exposures per second on ones, two seconds, the same 12 cm travel, the same six-exposure hold, and the same 90-degree turn. Route A moves the same distance on every exposure. Route B covers 9 cm in its first six exposures and 3 cm in its last six, decelerating into the pause. The prediction is that a decelerating approach makes a pause read as a decision and a full-speed arrival makes it read as an interruption; it may be wrong. When revising, change one thing, shoot B after A, keep A's source frames and sequence order, record settings, and consider shooting A, then B, then A again to check whether your setup is repeatable.

Further reading

• Capture a Small Time-Lapse Test for a Commercial Treatment — Read this when the treatment claim concerns elapsed time rather than posed movement: it draws the boundary between unposed interval capture and the stop-motion posing described here.
• The Motion Test Stutters in Preview. Check Playback Before Changing the Edit — If the playback lurches or the pause reads as a dropped frame, this checks preview delivery before you move any spacing, separating caching from the animation itself.