Give Treatment Words and Images Different Parts of the Explanation

A treatment page has two ways to look finished while saying very little.

The first is a handsome reference image with a paragraph of praise beside it. "Elegant, intimate, restrained." The reader can see that it is elegant. The praise has done no work that the picture was not already doing, and the page still hasn't said what the film does.

The second is a paragraph carrying the entire argument while the image sits there as wallpaper. If the words can be understood with the image removed, the image wasn't explaining anything. It was decorating a memo.

The correction is not a ratio. It isn't "sixty percent image, forty percent copy," and it isn't a rule that every noun appearing in both places is waste. The correction is to decide what each element is responsible for explaining, and then to let each one do its own job. An image demonstrates a visible relationship it genuinely contains: distance, scale, arrangement, light, texture, who is larger in the frame. The copy explains what a single frame cannot show: what moves, in what order, how long it takes, what it means, what should not be borrowed from the reference, and what is still undecided.

Name the decision this page must explain

Before choosing an image, write the page's question in one sentence. Not the concept. The question.

"How far apart do these two people stay, and who closes the gap?" "How much of this performance stays under the surface?" "How does the room get smaller as the scene proceeds?"

A page built around a real question has a natural test for every element. You can ask whether a given sentence or a given frame helps the reader answer it. A page built around a mood has no such test, so everything attractive gets in, and the copy drifts toward whatever adjectives the mood suggests. That is where "elegant, intimate, restrained" comes from. It isn't describing the film. It is describing the register that treatments are written in.

The question also tells you what the reader has to understand before they can judge your proposal at all. In a page about distance, that's the geometry of the space and the sequence of the gesture. In a page about restraint, it's where the performance is allowed to move and where it isn't. Whatever the reader needs in order to follow the argument is the page's actual content. Everything else is atmosphere, which the image may already be supplying.

The example running through the rest of this article is fictional, and the page described here is a plan rather than a built artifact: the still is described in words, not shown, and nobody has read the finished page.

The commercial: a person at a long table pushes a small handwritten invitation some way along the surface, stops with two fingers still on its near edge, waits, then lets go. The card comes to rest well short of the far end. The other person waits, then reaches the remaining distance and draws it in.

The page's question: How does this scene hold two people apart, and who completes the crossing?

Describe what the image really supplies

Look past the inventory. "A long table, a window, two people, an envelope" is a list of objects, and a list of objects is not yet a relationship. The relationship is the part that does explanatory work: the two people occupy opposite ends of the same long surface, the object between them is tiny against that span, and the frame gives one of them more room than the other.

That's what a still can carry, and it carries it well. Scale. Distance between bodies. Where the light falls and what it hides. Who is bigger in frame and who is small. What sits between two people and how much of it there is. A single frame can also carry a comparison, if both things are in the frame: him seated, her standing; the table's length against the card's width.

What a still cannot carry is nearly everything about time. A hand resting on a card is compatible with three different moments. The hand could be arriving. It could be pausing. It could be about to lift. The still does not tell you which, and no amount of looking will settle it, because duration isn't in the frame. For the same reason, a still cannot show that the hand kept two fingers on the edge while nothing else moved for four seconds. It cannot show intention, cause, or what happens next. It cannot show sound, or anything outside the crop. And it cannot show restraint, which is a judgment about a performance over time and not a property of a face.

A sequence relaxes this a little without dissolving it. Two frames can demonstrate that something changed between them. They still cannot tell the reader how long the change took, or whether the hand hovered. That information has to come from somewhere, and the somewhere is the copy.

One more distinction is worth keeping straight, because it changes what the page is allowed to claim. A frame from an existing film is a reference: it shows what some other scene did. A photo test you arrange yourself is your own construction: it shows a relationship you built and can defend. A diagram is a diagram. All three are legitimate, and all three are evidence of a different kind. None of them is the commercial, which does not exist yet. Whatever caption you write, using somebody else's frame is a permissions question that a credit line does not answer.

Make the words add the missing relationship

Here is the weak version of the page copy, sitting beside that wide still:

Elegant, intimate, restrained. Two people at a long table, the distance between them saying everything. A quiet, loaded moment.

Read it against the question. It answers nothing. "The distance between them saying everything" is a claim the image cannot support, since the image has no idea what the distance means to these people. The adjectives describe the register of the page, not the content of the film. And nothing in it tells the reader what happens after the frame.

Here is a revised version:

One locked wide, no cut. The card — a small handwritten invitation — starts at the near edge and stops about a third of the way along. Two fingers stay on the near edge. Nothing else moves for four seconds. Then the fingers lift and the card is left where it stopped, well short of the far end. He waits a beat longer than she does, then reaches the rest of the way and draws it in. The camera does not push in. The push-in from the reference does not transfer.

Now the division of labor is visible. The image establishes that these two bodies are far apart on one long surface and that the object between them is small. The copy supplies the movement, the held edge, the duration of the pause, the release, the fact that the card stops short, and the other person's response. That last one matters: without it, the page describes a gesture that goes nowhere.

Three things are happening in that paragraph that are easy to miss.

First, the copy is doing work no image could do, not work the image could do slightly worse. If a sentence could be replaced by looking harder at the picture, it's a candidate for cutting. If it couldn't, it's load-bearing.

Second, the copy is stating a departure from the reference. "The push-in does not transfer" is the kind of line pages usually omit, and its absence is expensive. A reference frame arrives as a bundle: composition, lens, movement, grade, performance, the whole apparatus of somebody else's film. The reader will assume the bundle travels unless you say otherwise. Naming which parts you're keeping and which you're dropping is not pedantry. It's the difference between a reference and an instruction.

Third, the copy is marking what it is. "Stops about a third of the way" is a proposal. "Nothing else moves for four seconds" is a proposal. Neither has happened. The page should read as though that's obvious, because a page that blurs proposal into recollection will be quoted back to you as though the scene were already shot.

What the revised copy does not do is remove description. The card is still "a small handwritten invitation" on the page, and it should be. Identification is not repetition. A reader looking at a white rectangle the size of a thumbnail cannot tell whether it's an invitation, a place card, or a napkin, and a reader who can't see the image at all needs that noun to follow anything. The thing to cut is mood restated as mood. The thing to keep is everything a reader needs in order to know what they're looking at and what it's for.

Change the image when the explanation cannot rescue it

Sometimes the copy is fine and the image is still arguing for a different film.

Suppose the still you love is a symmetrical two-shot: both figures equally weighted, centered, balanced, the table running straight down the middle of the frame. That is a beautiful image, and it is making a claim about equality. Your proposal is about imbalance, and about one person failing to complete a crossing that the other one finishes. The strongest thing in that frame contradicts the page, and no caption will fix it. Readers take the image's dominant behavior as the truth and treat the prose as commentary.

The options are limited, and pretending otherwise wastes everyone's time.

Replace it. Find a frame that carries the relation you actually want: one figure large and near, one small and far, the object between them, the asymmetry built into the composition rather than asserted in the copy.

Use a short sequence. Two or three frames can show the card at the near edge, partway along, and stopped short, with the second person's hand entering the last frame. A sequence can demonstrate that something changed. It cannot demonstrate how long the change took, so the copy still carries the pause. Sequences cost page space and attention; use one when the change is the point.

Make a diagram. A plan view of the table with two positions marked and the card's path drawn as a line is honest about being a diagram. It shows geometry cleanly and carries no atmosphere at all, and you shouldn't ask it to. A diagram that resolves the arrangement problem while the copy handles the performance may be the most accurate page available, especially at an early stage.

Label a partial reference as partial. If a single borrowed frame is the best you have for the space, say so on the page: "This is the distance, not the performance." One frame from another film does not demonstrate the whole scene, and a page that implies it does will be read as a promise you can't keep. Keeping the label short and specific is enough. The point is that a reader should never have to guess how much of the scene an image is standing in for.

Read the page through both routes

Once the page holds together, read it twice in two different ways.

The first reading is the argument: does the image and copy together describe one film? That means checking three states, and keeping them visibly separate on the page.

The right-hand column reflects where each piece of information is meant to come from in the plan. The real check comes when the page exists: read it and see whether the division actually held, or whether the copy crept back into describing what's visible while the action went unstated.

The second reading is the one available without the image. Not everyone will see it: some readers get the page as text, some will see it in a format where descriptions travel separately, and some will skim it in a meeting where the image is a thumbnail on a screen nobody can read. So the page needs a description of the still that reports what's in the frame — the long table, the two figures at either end, the small card under the near hand, the space between dominating the image. That description has a different job from the creative copy. It reports; the copy proposes. The two will repeat nouns, and that repetition is not a defect. One is for identification, the other for argument.

What a description should preserve is the same thing the image was carrying: the relationship, not just the inventory. "A table with two people" loses the point. "Two people at opposite ends of one long table, the card small between them, the distance filling the frame" keeps it.

Two cautions here.

Do not claim that concise copy makes a document accessible. Whether a given treatment is accessible depends on the format it's delivered in and how it was built: reading order, whether the description is programmatically associated with the image, contrast, how the file exports, whether the layout survives without the picture. A short caption doesn't settle any of that. Confirming it means checking the actual file, and that check hasn't happened for the page described here.

And keep the pair together. If the page gets reordered, exported, or handed to someone who swaps the image for a better one, the copy that carries the action has to travel with it. A still on one slide and its paragraph on the next is no longer a page with a division of labor. It's two half-arguments that happen to be filed near each other.

One page, two jobs

The revision this article has been describing is not shorter copy or bigger images. It's a change in allocation. Before, the image showed two people at a table and the copy restated the mood. After, the image establishes the spatial imbalance and the copy supplies everything the imbalance can't say by itself: the movement across the table, the fingers that don't lift, the four seconds, the release, the card that stops short, and the hand that comes the rest of the way.

One page, one question, and two elements each earning their place. That's the standard to hold the next page to. Not a ratio. Just a question you can state, an image that answers the part of it a frame can answer, and copy that answers the rest.

Frequently asked questions

What are the two ways a treatment page can look finished while saying little?

One is a handsome reference image with a paragraph of praise that only restates what the picture already shows. The other is a paragraph carrying the whole argument while the image is wallpaper. The correction is not a ratio; it is deciding what each element is responsible for explaining.

What can an image explain that copy cannot, and vice versa?

An image can demonstrate visible relationships it genuinely contains: distance, scale, arrangement, light, texture, who is larger in frame. Copy explains what a single frame cannot show: what moves, in what order, how long it takes, what it means, what should not be borrowed from the reference, and what is still undecided. A still cannot show duration, intention, cause, what happens next, sound, or restraint.

What is the first thing to write on the page?

The page's question in one sentence—not the concept. Examples include 'How far apart do these two people stay, and who closes the gap?' or, for the invitation scene, 'How does this scene hold two people apart, and who completes the crossing?' That question gives a test for every sentence and image.

What should be done when the image argues for a different film?

Replace it with a frame that carries the relation you want; use a short sequence if the change is the point, remembering the copy still carries duration; make a diagram if geometry is the issue; or label a partial reference as partial—for example, 'This is the distance, not the performance.' No caption fixes a dominant contradictory frame.

Does concise copy make a treatment document accessible?

No. Accessibility depends on the format delivered and how it was built: reading order, whether the description is programmatically associated with the image, contrast, how the file exports, and whether the layout survives without the picture. A short caption does not settle any of that, and checking it means checking the actual file.

Further reading

• When the Treatment’s Words and Images Argue for Different Films — Read this when the copy and the references imply incompatible but separately viable films and the conflict needs resolving before layout.
• Shorten a Treatment Without Cutting the Film Out of It — Turn here when a page must get shorter, to see which sentences carry movement, timing and proposal that no image can supply.