Quick answer

Keep one simple reference object in the same screen position in every shot, lock your camera distance and height, and change only the product's rotation or the background. Viewers then judge size against a constant, not against a moving frame.

A mug shot alone on a tabletop could be a thimble or a bucket. The camera gives no clue, because a close-up of a small object and a close-up of a large object look nearly identical. The usual fix — pulling the camera back to reveal more context — creates a second problem: the jump in distance reads as a cut, and continuity breaks.

The alternative is to stop moving the camera and let a fixed reference do the explaining.

Why a fixed reference beats a camera move

Scale is relative. Your viewer has no idea how far the lens is from the product, so apparent size carries almost no information on its own. What carries information is the ratio between the product and something the viewer already has a mental size for.

That is why a plain pencil next to a ceramic mug works. The pencil's length is familiar. If the mug's rim sits at roughly the same height as the pencil's midpoint, the viewer reads "small mug" without being told. If the mug dwarfs the pencil, the viewer reads "large mug."

A camera move can also reveal scale, but it costs you:

Continuity. A distance change between shots looks like a jump cut unless you motivate it.
Time. A pull-back needs several seconds to read clearly, which is expensive in a short clip.
Repeatability. Re-matching a distance by eye across a shoot day is fiddly.

A fixed reference costs you none of those. It stays put, so the only thing changing on screen is the product.

Step-by-step workflow

1. Choose one simple reference object and keep it in frame for every shot

Pick something with a shape and size most viewers already know, and something that will not compete with the product for attention. A plain pencil, a standard pen, a coin, a credit-card-sized card, or a sheet of paper all work. Avoid anything branded, anything with text, and anything whose size varies between versions.

Place it on the same surface as the product, slightly to one side, so it never overlaps the product's silhouette. Mark its position with a small piece of tape on the table if you are shooting across multiple setups — that mark is your anchor for the whole session.

2. Lock camera distance and height

Set the camera once, then do not touch the tripod. Distance and height together determine how large the reference appears, so if either changes, the reference stops being a constant.

Practical checks:

Tighten every tripod lock, including the pan and tilt drag.
Note the height on the tripod column, or tape the column so it cannot creep.
Use a fixed focal length if you can. Zooming changes apparent perspective in ways that are easy to miss on a small screen.
If your camera has stabilization that crops the frame, turn it off for locked-off shots so the crop does not shift between takes.

3. Change only the product's rotation or the background

This is the core constraint. Between shots, you may:

Rotate the product a quarter turn to show a new face.
Swap the background surface or backdrop.
Change the lighting direction, as long as the reference object stays lit consistently enough to compare.

You may not:

Move the camera closer or farther.
Change the lens or zoom.
Move the reference object.

In the mug-and-pencil example, shot one shows the mug's handle facing right. Shot two shows the same mug rotated a quarter turn so the handle faces the camera. The pencil has not moved a millimetre, and the camera has not moved at all. The viewer now sees two faces of the mug and still has the pencil to measure against.

4. Hold the final frame for about two seconds

Comparison takes time. If the clip cuts away the instant the rotated product appears, the viewer never gets to line up the reference against the new silhouette. Hold the last frame steady for roughly two seconds — long enough to compare, short enough not to feel stalled.

If your clip has more than two shots, hold each shot long enough that the reference object registers before the next cut.

5. Review by checking the reference object's apparent size

Before you publish, scrub through the clip frame by frame and watch only the reference object. Its height and width on screen should be identical in every shot, and its position should not drift. If it grows, shrinks, or slides, something moved — usually the tripod or the zoom.

Example input

Here is a copyable brief you can adapt. It describes a shot plan rather than a generated result; treat it as a starting point and adjust the wording to your own product.

text

Shot plan: two-shot product scale clip, locked-off camera.

Surface: plain matte tabletop, no pattern.
Reference object: one plain unbranded pencil, laid flat, tip pointing
toward the top-left of frame, positioned at the right edge of frame
so it never overlaps the product.
Product: unbranded ceramic mug, centred.

Camera: fixed distance, fixed height, fixed focal length, tripod
locked. No zoom, no push-in, no pull-back between shots.

Shot 1: mug handle facing right. Hold 2 seconds.
Shot 2: mug rotated a quarter turn so the handle faces camera.
Pencil and camera unchanged. Hold 2 seconds.

Lighting: single soft source from the left, unchanged between shots.
No text overlays, no logos, no on-screen graphics.

If you want to storyboard or previsualise the two frames before you shoot, you can build reference stills in the image workspace and then animate them — see AI image to video for that part of the pipeline. Keep the pencil in the same screen position in both stills, or the whole method falls apart at the storyboard stage.

Output review checklist

Run through this before publishing. The first three items are objective — you can verify them by scrubbing frames. The last two are judgement calls.

Check	How to verify	Type
Reference object size unchanged	Compare its on-screen height in first and last shot	Objective
Reference object position unchanged	Overlay or flip between frames; it should not slide	Objective
Product silhouette readable	Pause on each shot; can you tell what the object is and which way it faces?	Subjective
No text overlays, logos or invented UI	Watch the full clip once with sound off	Objective
Hold time sufficient	Count roughly two seconds on the final frame	Objective

If the reference object's size differs between shots, do not try to fix it in the edit by scaling one shot. That changes the product too, and the viewer's sense of scale becomes unreliable. Reshoot with the tripod locked.

Common mistakes

Moving the reference instead of the product. It is tempting to slide the pencil closer when you rotate the mug. Don't. The reference must be the one thing that never changes.

Letting the reference overlap the product. Overlap hides the silhouette you are trying to show, and it makes the size comparison harder to read. Keep a clear gap.

Using a reference with unknown size. A generic "small box" tells the viewer nothing. A pencil, a coin, or a sheet of paper carries a size most people already know.

Changing the background and the lighting at once. You can change either, but changing both makes the two shots feel like different scenes, which weakens the comparison. Change one variable at a time.

Cropping the reference out of frame. If the reference is only partly visible, viewers cannot judge it. Frame wide enough that the whole reference object sits inside the shot in every take.

Relying on a caption to explain size. Text overlays add clutter and often get skipped. The reference object does the job silently.

FAQ

Can I use more than one reference object?

You can, but one is usually clearer. Two references give the viewer two things to compare and can slow the read. If you do use two, keep both in the same positions across every shot.

What if my product is very large?

The same principle works, but your reference needs to be something the viewer can still see at that scale — a person's hand, a standard door, or a chair. The constraint is unchanged: the reference stays in the same screen position and the camera does not move.

Does the reference have to be a real object?

A real object is easier to keep consistent, because it cannot drift between takes. If you are building stills first, keep the reference's screen position and apparent size identical across those stills before you animate anything.

How long should the whole clip be?

Long enough to hold each shot for about two seconds. Two shots plus a short opening gives you a clip in the region of six to eight seconds, which is comfortable for a short-form format. Add shots only if each one earns its hold time.

What if I need to show the product from a different angle?

Rotate the product, not the camera. A quarter turn per shot is usually enough to reveal a new face while keeping the reference and the framing stable.

Can I fix a drifting reference in editing?

You can crop or scale a shot, but that also rescales the product, so the comparison stops being trustworthy. Locking the tripod during the shoot is the reliable fix.

The one rule to remember

The camera is the constant. The product is the variable. If you hold that split, viewers will read the size of your product correctly without a single word of explanation — and your shots will cut together cleanly because nothing about the frame moved.

How to Show a Product's Scale in a Short Clip Without Changing the Camera Distance