Pitch a Multi-Screen Film Without Pretending One Video Shows the Work
Pitch a Multi-Screen Film Without Pretending One Video Shows the Work
At 08:52 of a twenty-four-minute cycle, a figure stands up from a bench on screen A and walks out of the frame. At 08:52, a shadow begins crossing the lit floor of screen B, thrown by something standing outside B's picture. It takes four seconds to reach the doorway at the far end of B's corridor. At 08:57, while screen C still shows its own doorway empty, a latch turns on the channel beside screen C. C's doorway stays empty until 09:04.
No single video file shows a curator what that paragraph shows. That isn't a production problem you can edit your way out of. It's the work.
The piece I'm describing is invented — a fictional three-screen proposal called Neither Side, built for this article so that the comparisons stay checkable. Its numbers don't move: three screens, three sound channels, a portable loop with no designated start, a forty-second interval indexed at 08:40–09:20 from an arbitrary point in the cycle. What follows is a method for choosing pitch evidence when the thing you're pitching only exists as relationships.
Find the relation that disappears on one screen
Before you choose a single asset, write down what a single-channel extract would lose. Not in general. In this passage.
For Neither Side, five things go. The coincidence at 08:52, which is the entire proposition — someone leaves one side of a closed passage while something arrives on the other, and the piece refuses to say whether they're the same person. The disagreement between C's empty image and C's latch. Scale: A is four metres wide and shows the waiting room at roughly life size; C is a metre and a half and shows its doorway at about 0.4 scale, so a person crossing C's threshold would read at less than half the height of the person on A. Direction: at the doorway position the latch arrives from beside your own shoulder, not from ahead. And position: no place in the room holds all three images legibly at once.
That last one is a design fact, not a flaw to hide. Room is eight metres deep, six across. A sits on the north wall, four metres wide, facing into the room. B is on the east wall, two and a half metres, facing west. C is mounted low on the south wall, a metre and a half, facing north — the same wall as the entry door, near its far end. Stand on the threshold and A is face-on eight metres ahead, B is at sixty or seventy degrees to your right, and C presents only its edge. Walk to the centre and turn north-east and A and B both sit at about forty-five degrees, legible but never clean, with C behind you. Turn south to watch C's empty doorway and A goes behind you instead.
So the relation that must not disappear is simultaneity under a viewing cost. Write that sentence at the top of the pitch. Not "an immersive three-channel installation exploring waiting." The equipment list and the mood word are the two easiest ways to bury the actual proposition, because both are true and neither constrains anything.
Three evidence objects, three different failures
Once you know the loss, you can ask which object covers it. Each of the three standard options carries one dimension and quietly drops the rest.
A single-channel excerpt. Take three minutes of A's material, sum B's and C's audio into stereo, and you have a real demonstration of image texture, pacing, and the quality of the waiting — the drape of the light on the bench, how long the figure sits before standing. What you no longer have is the relation. B's shadow never crosses that frame, and C's empty doorway is absent with it; the 08:52 coincidence isn't misread in the excerpt, it's unshown. Nothing in it tells a viewer that anything is happening anywhere else at that moment. The stereo fold is worse. It doesn't omit; it asserts. The latch, which at the threshold position reaches you from your own right rear, now arrives in both ears from the front. Your excerpt states the opposite of the room.
A synchronized three-track timeline. Rows for A, B, C, an axis for the forty-second interval, the 08:52 coincidence visible as a vertical line through two rows. This is the only object that makes simultaneity explicit rather than implied, and the only one where the empty doorway and the latch can be read at the same x-position and seen to contradict each other. Its limits are equally plain. A timeline has nowhere to put four metres against one and a half, nowhere to put the sixty-degree angle from the threshold, and no way to express "this sound comes from behind you." It also smuggles in a falsehood unless you annotate it: an axis starting at zero implies a beginning the work doesn't have. Mark the axis as indexed from an arbitrary point in a continuous loop, or the curator will ask when the piece starts, and you'll have taught them that it does.
A plan-view sketch. The room to scale, three screens, the entry, two marked positions, speaker placement beside each screen, sightline arrows from each position with the angle noted. This is the only object that answers the curator's actual question — what will it be like to be in there — and it makes the viewing cost discussable in the room rather than in a paragraph. What it cannot do is duration. It has no axis for time, so the 08:52 relation and the empty-doorway contradiction simply aren't in it.
None of these wins. The choice is a combination matched to the decision the recipient has to make. If the curator is asking whether there's enough to hold attention, the excerpt does more work than the sketch. If the question is how this differs from a single-screen version, the timeline does. If it's whether the piece will work in a specific gallery, the sketch does, and the other two are supporting documents. More assets are not more explanatory; a pitch with all three and no stated loss just asks the curator to do your comparison for you.
A visitor who arrives in the middle
The piece has no fixed beginning. Someone walking in at 08:20 hears channel C working before they see any image, because the speaker beside screen C is on the wall they just came through. They're on the threshold. A is clear and large and already occupied — a figure on a bench, a closed door. B is a smear at the right edge of vision, a corridor with light on the floor. C is an edge-on panel they have to walk past to see at all. They stay, because A is doing something and because the latch keeps arriving from the wrong direction.
Then the shadow crosses B and finishes at B's doorway, and a second later the latch turns on the channel beside C — and the doorway on C, which they still haven't seen, is empty. To check it they have to cross the room and turn around, which costs them A entirely. That's the reason to move, and it's also the reason the piece can't be pitched as a fixed programme.
Notice what this description is: a set of positions and their consequences, not a path. You don't know where a visitor will stand, how long they'll stay, or whether they'll turn at all. Neither do I, and neither does the curator. Say "from the threshold position, C is not visible" rather than "visitors see A first, then turn to C." The first is a fact about the room. The second is a guess about strangers that the pitch will be graded against later.
If your piece does have a fixed beginning, this section changes shape, and mostly gets easier: you can lean on sequence, and a linear excerpt becomes more defensible because the work itself is linear in the viewer's experience. The mistake to avoid is borrowing that licence when your work doesn't have it. And whatever the case, don't claim that every position delivers the same information. In Neither Side they demonstrably don't, and a curator who visits will find that out.
What the demonstration can't settle
The planned video study is forty seconds, one fixed camera on the threshold position, three channels mixed down with C panned to the camera's rear right. Its most useful function is not to show the work. It's to show what a position costs. In frame: A's figure leaving, B's shadow at the extreme edge, partly cut off. Out of frame: C's image, entirely. A curator watching it learns that A and B can coincide and that this particular viewpoint wastes B — which is exactly the argument the pitch is making.
Label it that way in the document. "Fixed camera, threshold position, forty seconds, channels mixed to stereo with channel C placed to the camera's rear right. Screen C is not visible in this study." State the substitution. If the study were ever to stand in for the piece without that label, it becomes a false claim about the work rather than a demonstration of one viewpoint.
Then separate two lists that makers routinely blend. Requirements are artistic: A and B need to be visible together from somewhere near the centre; C's speaker must not be co-located with A's; the room must be dark enough that B's shadow reads, and long enough that the threshold and centre positions are genuinely different places. Capabilities belong to the venue and its technicians: ceiling height, load-in, acoustic separation, blackout, whether the south wall can take a low-mounted panel at all. You can assert the first list. You cannot assert the second until someone with the building has answered, and the pitch should say which list each line belongs to.
For a sense of how much coordination a spatial encounter involves, MoMA's account of installing Isaac Julien's Ten Thousand Waves (Inside/Out, November 25, 2013) describes a nine-channel work with multiple viewing positions and coordinated screen and sound placement — which is enough to make the point that the spatial arrangement is authored, not incidental. It's an institutional first-person account of one historical installation, and it says nothing directly about your room. It's evidence that coordination is the norm at this scale, not a specification for your piece, and a curator will read it that way.
The smallest set that makes the proposition assessable
Start with a relation statement: one paragraph naming what must coexist and what the visitor's position does to it. In Neither Side, that's the 08:52 coincidence, the empty doorway under the latch, and the fact that no position holds all three. Everything else in the package either supports that paragraph or is decoration.
Then the two objects that complement each other cleanly. The three-track timeline carries the forty-second interval and the full-cycle bar beside it, with the axis marked as arbitrarily indexed. The plan view carries the room, the screens, the speakers, the two positions and their angles. Put them on facing pages or side by side, not in sequence. A pitch document reads left to right, top to bottom, and that reading order is itself a claim — it says these things happened one after another. Simultaneous material laid out in a vertical list of paragraphs contradicts itself. Set the simultaneous things beside each other and let the page stop arguing with the work.
Add the excerpt only if it adds something the recipient doesn't already have. If the curator has seen previous work from you, the image texture may be established and the three minutes may cost you the attention you need for the relation. If it does go in, caption it the way you caption the study: which channel, what was folded, what was dropped. And keep the two optional pieces genuinely optional. A package whose smallest useful version is one page plus one diagram is stronger than one that needs four attachments to make sense, because the curator will forward the four attachments and the one page will arrive alone.
Then test it before you send anything. Hand the package to someone who hasn't heard the piece described and ask them to say back what has to be simultaneous and what's still undecided about the room. If they describe it as a three-screen video about waiting, the relation statement isn't doing its job yet. If they tell you which position loses C, it is.
Also check the objects against each other. A sketch showing two speakers and a timeline showing three channels is a credibility leak, and curators read diagrams closely because diagrams are where makers stop hedging. Count the channels in every object. Confirm the interval, the screen sizes, and the position labels agree. Resolve that before you spend a single hour on the visual polish of the package, because polish applied over a contradiction makes the contradiction look deliberate.
Say what the pitch cannot carry
The combination to send: the relation statement, the annotated three-track timeline over the forty-second interval, and the plan view with two positions and speaker placement. The threshold study, if it gets made, goes in as a labelled demonstration of a viewpoint's cost. The single-channel excerpt goes in only if the recipient needs image and sound texture, and it arrives with its channel and its stereo fold written on it.
Then say the limit out loud, in the pitch, in plain language. These documents cannot reproduce standing in that room, four metres of screen resolving into a face, or a latch arriving from behind your right shoulder while the doorway in front of you stays empty. Nobody can reproduce that in a file. What the package can do is make the proposition checkable: a curator can read exactly what has to happen at the same time, see where a visitor has to stand to catch it, and tell you which parts of the room are still a promise rather than a plan.
That's the whole job. Not proof. Assessability — and an honest account of the gap the room is going to fill.
Frequently asked questions
Why can't a single video file show a multi-screen film like Neither Side?
No single video shows what the relation paragraph shows. A single-channel extract loses the 08:52 coincidence, the disagreement between C's empty image and C's latch, scale differences, direction such as the latch arriving from beside the viewer's shoulder, and position because no place in the room holds all three images legibly at once. The stereo fold is worse than omission: it asserts that the latch arrives in both ears from the front, the opposite of the room.
What should be written before choosing pitch assets?
Write down what a single-channel extract would lose in this specific passage. For Neither Side, the relation that must not disappear is simultaneity under a viewing cost. That sentence belongs at the top of the pitch, not an equipment list or a mood phrase like an immersive three-channel installation exploring waiting.
What do the three evidence objects each cover and drop?
A single-channel excerpt can demonstrate image texture, pacing, and the quality of the waiting, but drops the relation and misstates direction through the stereo fold. A synchronized three-track timeline makes simultaneity explicit and can show the empty doorway contradicting the latch, but has nowhere to put scale, viewing angles, or sound coming from behind; its axis must be marked as arbitrarily indexed in a continuous loop. A plan-view sketch answers what it will be like to be in the room and makes viewing cost discussable, but cannot show duration or the 08:52 relation.
What should be tested before sending a multi-screen pitch package?
Check the objects against each other: count the channels in every object; confirm the interval, screen sizes, and position labels agree. A sketch showing two speakers while the timeline shows three channels is a credibility leak. Also hand the package to someone who has not heard the piece described and ask them to say back what has to be simultaneous and what is still undecided about the room. If they describe it as a three-screen video about waiting, the relation statement is not doing its job.
What can the pitch package not do, and what should it say about that?
It cannot reproduce standing in the room, four metres of screen resolving into a face, or a latch arriving from behind your right shoulder while the doorway in front of you stays empty. The package can make the proposition checkable: a curator can read what has to happen at the same time, see where a visitor must stand, and tell which parts of the room are still a promise. Any demonstration should be labeled with its substitution, and requirements should be separated from capabilities.