Make a Small Anamorphic Artwork Test for a Commercial Pitch
Make a Small Anamorphic Artwork Test for a Commercial Pitch
An anamorphic artwork is not one image. It is a set of marks scattered across surfaces that don't agree with each other, plus one viewing position that resolves them. From that position, the ring looks round. From anywhere else you can see what is actually on the table: a long, stretched smear on the horizontal, a short arc standing up on a card, meeting along a seam that neither of them respects.
That is the whole mechanism, and it is exactly what a flat mockup conceals. In a graphics application you place the mark, choose the camera, and look. Nothing stops you from swinging the camera a few degrees, and nothing tells you what the swing cost. The mockup has no opinion about where the camera was, because it was never anywhere.
So the answer to "how do I test an anamorphic idea for a pitch" is a small physical build, not a better render. Choose a stable set of tabletop surfaces and a viewing point that matters to the shot. Fix the design from that point with safely supported projection. Transfer the design into removable physical marks. Then turn the projector off and photograph the marks twice: once from the recorded point, once from a deliberate offset. The pair is the evidence. One flattering photograph of the aligned view is not.
Nothing in the worked example below has been built. No projector has been set up, no marks applied, no photographs taken, no camera move tested. It is a specification you can run in an afternoon, written so that the result means something whichever way it comes out.
Design the image and the viewing relationship together
Pick the design and the viewpoint in the same decision, because they are not two decisions. The surfaces only receive a pattern; the viewpoint is what turns that pattern into a picture.
For a first test, keep the shape to something with one measurable feature. An open ring works well: a circle with a single gap in it. The gap gives you something to check that is more precise than "does it look round" — the two ends either face each other across the break, or they don't.
Set the ring so that the seam between the card and the tabletop cuts through it. Everything below the seam in your design lands on the flat base; everything above lands on the upright card. Because the ring crosses the seam, it crosses it twice, once on each side. Make one crossing continuous, a single stroke that runs off the card and onto the table without pausing. Put the gap at the other crossing, so that one end of the ring stops just above the seam and the other stops just below it.
Now you have two registration witnesses on two different planes. If the surfaces are where you think they are, the gap reads as a clean break and the continuous crossing reads as an unbroken curve. If they aren't, the gap's ends slide apart or overlap and the continuous crossing shows a kink. You will not have to squint to judge it.
The proportions matter less than the relationship, but here is one workable set, offered as an example rather than a formula: a card standing about 60 cm high on a tabletop around 80 cm deep, with the projector at roughly your own eye height and a metre or so back from the seam. Scale that up or down to suit the room. What matters is that the whole design lands on the surfaces with margin to spare, because a mark that runs off the edge of the card cannot be checked.
Decide the viewing position before you trace anything, and write it down in numbers rather than remembering it. Three quantities are enough: the height of the lens above the tabletop, its horizontal distance back from the seam, and the direction it faces. Add a tape mark on the floor under the lens and another on the table, and mark the position of the card's bottom edge on the table as well. If the card shifts by a centimetre, the marks you already drew become wrong, and you need to know that the shift happened rather than discover it later in a photograph you can't explain.
One caution about ambition. Thomas Quinn's first-person account of making anamorphic illusions describes projecting a design onto intersecting surfaces and tracing it from a fixed projection position, which then becomes the position the picture coheres from (Medium). That is the mechanism this test uses, and it is the only part of the account being borrowed. His described mounting involves height and improvised support; this version stays at table level with the projector on a stable surface and nothing clamped to anything that isn't rated for it. If you can't reach the intended viewpoint from a tabletop setup, make the design smaller rather than the ladder taller. Draw your own shape, too — an anamorphic test travels to meetings, and a borrowed character traced onto foam board is a problem you don't need.
Project, trace, and preserve the setup
Fix the surfaces first. The card must not bow, lean, or creep, and the base must not slide. Tape the card's base down, brace it if it wobbles, and check it by pressing gently before you start. Everything that follows assumes the geometry is frozen. If it isn't frozen, you are measuring the wobble, not the idea.
Project your design from the recorded position. Dim the room enough that the projected edge is visible, focus it, and let it settle. Then trace.
Use marks you can remove. Chalk or a white marker on a dark surface, a pencil on light card, low-tack tape or removable vinyl if you want the lines to have width you can see from across a room. Removable matters for two reasons: you'll want to iterate, and you'll want the surfaces back afterward. Test your marking tool on an offcut first. A projector in a lit room is dim, and it is easy to trace a pale line that seemed obvious under the beam and turns out to be invisible once the light is off. If the marks can't be read without the projection, the test can't demonstrate anything.
Trace deliberately. Follow the outer and inner edge of each stroke if you want the ring to have thickness. Mark both ends of the gap with a short cross-tick so you can find them again. Mark where the seam falls inside the design, and put small registration ticks at the outer corners of the projected frame so the whole thing can be re-found if something moves. Keep your hand out of the beam while you work, and work on the far side of the surface you're marking so you don't block your own reference.
Before you touch the projector, record the arrangement. Photograph the setup from off to the side — a view that shows the card, the table, and the projector in position, so that someone who wasn't there can see what made the marks. That is your setup diagram, and it does a different job from the two inspection shots later. Write the three measurements down next to it.
Then move the projector. The marks stay. This is the step that separates a physical anamorphic artwork from a projection-mapping preview. If the artifact is light, switching the projector off destroys it, and what you've built is a mapping test, not an object that carries its own distortion. That's a legitimate thing to build, but it is a different construction and it answers a different question. A photograph of projected light cannot stand in for an artwork that exists when the light is gone.
One last thing before you strike the setup: the fact that the projector was on the same support doesn't guarantee that the marks are registered. Card stock warms and moves. A tripod head sags a millimetre when you release it. Tracing has width. Expect small errors, and go looking for them rather than assuming they aren't there.
Inspect the physical marks from both views
Projector off. Room light whatever makes the marks read best.
Put a camera back at the recorded position — as close as you can get to it. The camera cannot occupy the point while the projector is standing in it, which is why you recorded the position before moving anything. "At the position" means the lens, near enough: lens height above the tabletop matching the recorded height, lens distance back from the seam matching the recorded distance, and the same heading. Use the same focal length for both shots in the pair, because changing the lens changes the composition and destroys the comparison. Support the camera; don't handhold either frame.
Now check the aligned view against the reference. The reference is your original design as a flat image, the composition you intended someone to see. Print it small, hold it beside the shot at matched size, or overlay it at partial opacity. Decide the tolerance before you look — how far can the gap's two ends be from facing each other and still count as a match? If you decide afterward, you will rationalize whatever came back.
A mismatch here is not a failure of the idea. The aligned view tests your build. If the gap's ends miss, look for the ordinary causes: the card moved, the projector was nudged, the camera isn't actually at the recorded point, or the traced line has enough width to shift the apparent join. Document the mismatch in the frame, don't crop it out.
Then move off axis, deliberately and by a known amount. Step the camera sideways along the table. Then try a height change. Then a straight push in or back along the sight line. Each one fails differently and the differences are the useful part. A lateral move slides the arc on the card sideways against the arc on the table, and the gap's ends separate circumferentially — the exact thing the design was built to detect. A height change re-foreshortens the flat arc and the ring goes fat or thin. A change of distance along the sight line is usually the most forgiving of the three, because everything scales together; the shear survives, even though the perspective doesn't perfectly. That last point is worth testing rather than assuming, because if it holds in your build, a dolly along the axis is cheaper to shoot than a lateral track, and that is a practical thing for a director to know.
Record each offset the same way you recorded the aligned position: distance moved, direction, and height. Two photographs without the numbers are two anecdotes.
Be clear with yourself about what you have at the end of this section: two still frames. Not a move. If the shot you're proposing involves the camera travelling, you have not tested it, and the two stills don't add up to a movement test.
Decide what the commercial's camera can do
Now connect the observation to the shot on the table. Commercial camera proposals for an anamorphic image generally fall into three shapes, and the small test supports them very unequally.
A held aligned view. The camera sits at the recorded point and the image reads correctly for the duration of the shot. This is the lowest-risk proposal and the one the small test supports best. What it demands is boring and expensive: a mounting that puts the lens at the recorded point, on a surface rigid enough that nothing creeps between takes, and a set that holds registration for however long the shot runs.
A reveal from misalignment. Begin off axis, end on it, and watch the scattered marks resolve into the design. This is the second comparison the small test proposes and has not performed. If you want to pitch it, build it as a camera move — start at a recorded off-axis position, end at the recorded aligned position, and shoot it as motion rather than two stills. The question the pitch has to answer isn't whether the reveal works; it's whether the moment of resolution lands on the product moment. A reveal that finishes two seconds early spends its payoff on an empty frame.
A move that must preserve coherence. Any tracking, arcing, or crane move that keeps the image correct throughout. The small test cannot speak to this at all. Coherence through a move is a continuous problem and needs continuous checking, not two sample frames. Say so in the room, and if someone wants the move, price a movement test as its own piece of work rather than folding it into the small study.
Whatever you propose, be precise about the boundary of what the tabletop test proves. It demonstrates the mechanism, it shows the viewpoint dependence in a way a flat mockup never will, and it sizes the tolerance in one small case. It does not establish that the design will work at architectural scale, that the surface on location is rigid enough, that the projector can throw far enough in the real room, that you can get permission to install anything anywhere, or that an unrestricted camera move will hold. Those are separate questions with separate answers, and a client who has been shown two honest photographs is far more likely to accept that than a client who has been shown one beautiful one.
What goes to the pitch
The deliverable is three things, not one. The original design as a flat image, showing what the marks are supposed to resolve into. The aligned view, showing what actually happened. The offset view, showing what the whole thing costs when the camera isn't where it's supposed to be. Put the setup photograph and the recorded positions beside them so the geometry can be understood and the test repeated.
Then say which of the three camera shapes you're proposing, and say it in one sentence. Hold the alignment, reveal it, or commission a movement test first. If you're holding or revealing, the boundary on the claim is the test's own boundary, and you can state it in the same breath as the proposal.
Keep the off-axis image in the explanation. It is tempting to lead with the aligned view, because it's the one that looks like an idea, and the offset view looks like something went wrong. But the offset view is not a mistake being confessed. It's the mechanism made visible — the same marks, the same surfaces, a camera in the wrong place. It tells the room where the illusion lives, and it's the reason to believe the aligned frame wasn't the only one you could produce. A pitch that shows only the good angle is asking everyone to take on trust the one property of anamorphic artwork that the whole design depends on.
The gap in the ring is where this gets decided. Two ends, facing each other across a break, on two planes that don't agree — until the camera stands where the projector stood. Move it and the ends tell you. That's a small thing to build, and it's a better argument than a render.
Frequently asked questions
What is an anamorphic artwork, and why does it need more than one viewing angle?
It is a set of marks scattered across surfaces that do not agree with each other, plus one viewing position that resolves them. From that position the ring looks round; from elsewhere you see a stretched smear on the horizontal and a short arc on a card, meeting along a seam. The offset view shows the mechanism.
Why is a flat mockup insufficient for testing an anamorphic idea?
In a graphics application you place the mark, choose the camera, and look; nothing stops you swinging the camera, and nothing tells you what the swing cost. The mockup has no opinion about where the camera was, because it was never anywhere.
What is the basic small physical test?
Choose stable tabletop surfaces and a viewing point that matters to the shot. Fix the design from that point with safely supported projection. Transfer the design into removable physical marks. Then turn the projector off and photograph the marks twice: once from the recorded point and once from a deliberate offset. The pair is the evidence, not one flattering aligned photograph.
What does turning the projector off separate?
It separates a physical anamorphic artwork from a projection-mapping preview. If the artifact is light, switching the projector off destroys it, and a photograph of projected light cannot stand in for an artwork that exists when the light is gone.
How does the small test support the three camera proposals, and what should go to the pitch?
It supports a held aligned view best, the lowest-risk proposal. A reveal from misalignment needs a movement test shot as motion, not two stills. A move that preserves coherence needs continuous checking and cannot be addressed by the small test. The pitch should include the original flat design, the aligned view, the offset view, the setup photograph, and the recorded positions; state which camera shape is proposed. The test does not prove architectural scale, location rigidity, projector throw, permission to install, or that an unrestricted camera move will hold.