Some Commercial Versions Need Different Filming, Not Better Cropping

A master can be cropped, selected, retimed, reordered, reframed, stabilised and regraded. What it cannot do is hold a relationship the frame never contained at a usable scale, or stretch a two-second hold into a setup. So the useful question in a treatment is not "can we cover this with one master?" It is "what does each version need from the same action, and does the master actually contain it?"

That question is easy to answer badly, because it invites a yes. A director says the wide covers everything; a producer asks what the extra setup is for; the treatment gestures at "flexibility." What follows is one deliberately invented scene worked through three ways, then translated into the sentences a treatment can actually carry. Nothing here was shot, timed, scheduled or priced. It is a paper comparison, and it is most useful read that way.

Put the competing requirements on the same action

Take the version list as the brief describes it and reduce it to what each version must preserve. Not the file, the frame, or the duration: the action. Then ask three things of that action. How long does it take? Which spatial relationship carries it? What must the viewer already know before the payoff? Answer all three for every version before deciding anything, because a version that fails on shape and a version that fails on duration need different remedies, and a treatment that blurs them proposes the wrong one.

The invented scene. A long table, shot side-on. A giver at one end, a recipient nearer the camera, and a colleague seated just outside the frame beyond the recipient. The giver pushes a parcel; it slides the length of the table and stops within the recipient's reach. The recipient, still turned toward the colleague off-frame, does not look down. Time passes with the parcel unclaimed. Then the recipient notices it, draws it in and turns it over. That wait is the misunderstanding — the parcel sits there long enough that the audience reads it as passing through rather than arriving, and the discovery at the end corrects them.

The brief asks for a long landscape film and a short portrait execution. Under that stipulation:

The long version needs duration, and the master has it. It needs the traverse — the parcel crossing the table's length is the shot's main event and the source of its distance. It needs the wait before the turn. Landscape holds all of it: the giver at one end, the table's width across the frame, the slide, the recipient, the unclaimed interval.

The short version needs a setup that explains why the discovery is a discovery, and it needs it in whatever fraction of the time the long version had. Its constraint is duration first. Its constraint is shape second: a portrait frame centred on the recipient keeps the recipient and the parcel, but the giver and the far end of the traverse fall outside it, and no vertical composition holds both ends of a long table seen side-on.

Note what is not broken. Information order survives. The parcel arrives before it is claimed, the claim precedes the discovery, and the master records that sequence cleanly. A short version lifted from the take keeps the order and loses the time. Order is the one of the three constraints this scene doesn't have to solve, and saying so is part of the analysis — it keeps the treatment from proposing a re-edit for a problem that editing cannot touch.

Separate a post-production operation from absent material

Once the requirements are laid out on the action, classify the shortfall by what the master would contain.

Some problems are selection. The take runs twenty seconds; the short version wants to open with the parcel already at rest. Start later in the take and you have it — the parcel sits, the recipient is mid-sentence. Post can do that, and the treatment should say so plainly rather than treat it as a coverage question.

Some problems are composition. The giver's waiting posture lives in the landscape frame at the far end of a long table. A portrait crop around the recipient will not include it, because the distance that gave the shot its meaning is the same distance that puts the giver outside the vertical frame. Post cannot reorganize its way out of that. Cropping is subtraction. It moves you nearer to what is there; it does not add back what the composition left out.

Some problems are not in the material at all. The colleague's reaction to the parcel arriving beside them is outside the frame and stays outside it. And if any version's payoff were the recipient's face after understanding — a look up, a change of expression — that beat was never performed, because the scene as staged ends on the parcel being taken. A bigger negative, a larger format, a wider master: none of them produce a performance that did not happen. Higher resolution makes existing information cleaner. It does not make a distant figure's expression legible if there was never enough of it in the frame to survive the extraction, and it certainly doesn't manufacture a beat.

That last point deserves care, because it is easy to assert. Whether the giver, at the far end of this table, reads at all is a question about a specific frame — the figure's size in it, the lighting, the angle of the head. The honest position is that the extraction would very likely deliver scale without much performance, and that this is worth checking rather than assuming. Assuming it is exactly the reflex that produces a treatment promising a portrait version and a landscape version from one setup, followed by an edit that discovers otherwise.

The general rule: describe the missing moment precisely. "The giver's reaction" is a wish. "The giver's face, at the far end, at roughly a twentieth of frame height" is something a producer or a cinematographer can answer.

Compare three credible routes

Route one: the same action, staged less demandingly. Instead of the side-on wide, put the camera at the table's end and shoot down its length. The recipient sits near the lens, the giver further back, the parcel travelling toward camera. Now the separation between the two people runs in depth, and a vertical frame holds both of them, because vertical frames are good at depth and bad at width. Compress the wait at the same time: the recipient's inattention has to be visible immediately, the parcel arriving beside a hand that does not move toward it, the turn arriving sooner.

What this buys is one staging that serves every version. What it costs is the long version's geometry. The traverse across the table's length was doing expressive work — the distance, the far-end giver, the sense that the parcel had a journey. A depth staging replaces that with a corridor. Depth also brings its own problems: the near shoulder can occlude the far figure, and a background giver may read as furniture rather than as a participant. And a shared route changes the long version, which had no problem to solve. Whether the compressed wait still reads as a misunderstanding in the short version, and whether the depth staging still convinces in the long one, are both questions that only a read of the material can settle. They are worth stating as questions.

Route two: keep the master, add the minimum that the other version needs. The landscape master stays as staged. One additional vertical setup at the recipient's end carries the parcel arriving into a frame that holds the recipient, the near table edge and the discovery, with the far end implied rather than shown. That setup exists for one reason: the wait, which is the short version's setup, cannot be recovered by cropping, and cannot be lengthened into a running time the short version has to spend elsewhere.

The dependency is not the camera move. It is a second pass — the parcel in the same state, arriving at the same place, performed again for a different frame — plus the reset of the prop's position between takes. Once the action has been shot for the landscape frame, its end state has already happened; shooting it again for portrait means resetting the table and running it twice.

Two cautions about added material. It is not automatically sufficient: a portrait setup that simply repeats the wide's information at closer range does nothing for the short version, which doesn't need a tighter version of the relationship, it needs the wait carried in a form the vertical frame can hold. And an insert of the parcel alone doesn't do it either — a parcel sitting on a table establishes a parcel on a table, not a giver waiting and a recipient not noticing. The added material earns its place only if it supplies the specific thing that went missing.

It is also not automatically free. AICP's 2026 Bidding Reference Guide, in the pages covering bid expectations and deliverables and post production, puts the question of how extra footage or extra takes support different frames and cutdowns into the bid conversation. That is the whole of what the guide is doing here, and it is worth being exact about its weight: it is US production-association guidance written in question form, it sets no rate, it establishes no entitlement or approval, and it confirms nothing about whether a particular additional setup is feasible. The reading it comes from is dated 8 September 2026 and has not been renewed for this piece. It is good evidence that the question belongs in a bid. It is not evidence about this scene.

Route three: reduce a version's ambition, with approval. The portrait execution stops trying to carry the misunderstanding. It opens at the moment of noticing and ends on the turn: the parcel, the recipient's hand, the discovery. The wait is gone, and with it the need for any capture beyond the landscape master.

The cost is creative, not financial, and should be named that way when no estimate exists. The two versions stop saying the same thing. If the campaign's idea depends on the misunderstanding — on being wrong about a parcel for a beat and then being corrected — the short version no longer carries the idea. It carries a moment from it. That may be entirely acceptable for a portrait placement, and the person who owns the brief is the person who says so. Reducing a deliverable's ambition is a request, not a production decision and not an editor's.

Make the choice visible without writing the bid

A version matrix is useful when it ends in a choice and useless when it lists every shot anyone has imagined. Keep it to the columns that force a decision: what the version must preserve, what the shared master supplies, what is missing, what that depends on, and whose call it is.

The treatment sentence that comes out of this is short. Something like: the landscape master carries the pass, the wait and the turn, all three inside the frame and inside the long version's running time. For the portrait execution we propose one additional vertical setup at the recipient's end, as a stated exception to the shared-master plan and for one reason only — the wait cannot be cropped out of a landscape frame, and cannot be given a running time the short version does not have.

Then say what is unresolved. Here it is whether the short version is permitted to communicate something different from the long one. That single answer decides whether the additional setup is necessary at all, and it is not a question the treatment can settle by writing more of itself.

Everything downstream belongs to other people, described as questions rather than answers: the schedule and cost of a second pass and a prop reset, the technical check of whether a portrait extraction from the wide holds the recipient and the parcel at usable scale, and the approval for any change to the deliverables. A treatment that names the creative dependency and routes the rest has done its job. One that quietly promises savings, or assumes approval, has not.

Where this leaves the proposal

The route worth proposing is the second one: keep the landscape staging the long version needs, add one vertical setup for the portrait execution, and be explicit that it exists because of a specific absence rather than a general appetite for coverage. Success here is a reasoned capture proposal with an exception in it — not a promise to shoot everything, and not a claim that one master serves every version.

Two facts would change the recommendation, and both are cheap to establish. If the brief owner confirms that the portrait execution only needs the arrival and the discovery, route three removes the additional setup entirely, and the price is a change in what the short version says rather than anything on a call sheet. If a look at the landscape frame shows that a portrait extraction holds the recipient, the parcel and enough of the far end to keep the relationship alive, the additional setup may be unnecessary — the composition problem would have been smaller than assumed. Neither has been tested. Both are the kind of question a producer can answer in an afternoon, and both are better asked before the bid than after the edit.

What the treatment is entitled to conclude is narrow and useful: the master covers the long landscape version, it does not cover the short portrait version's setup, and the options are a shared staging with a creative loss for the long version, one additional setup with its own dependencies, or a shorter ambition for the short version. Which of those is affordable, schedulable and permitted is not the writer's determination. Saying so clearly is.

Frequently asked questions

Why can't a landscape master always be cropped into a portrait version?

Cropping is subtraction: it can move nearer to what is in the frame, but it cannot add back what the composition left out. In the side-on table scene, a portrait frame around the recipient loses the giver and the far end of the traverse. The body distinguishes that from selection problems, which post can solve by starting later in the take.

What three routes does the piece compare for the short portrait execution?

One is shared staging: shoot down the table's length so a vertical frame holds both people, and compress the wait. Another keeps the landscape master and adds one vertical setup at the recipient's end. The third reduces the short version's ambition, opening at the noticing and ending on the turn, with approval because it no longer carries the misunderstanding.

What must added coverage supply to be worth shooting?

It must supply the specific thing that went missing, not merely a tighter repeat of the wide or an insert of the parcel alone. For the portrait version, the missing element is the wait carried in a vertical frame; a parcel insert establishes only a parcel on a table, not a giver waiting and a recipient not noticing. Added material also brings a second-pass and prop-reset dependency.

What creative question can decide whether an extra setup is necessary at all?

Whether the short version is permitted to communicate something different from the long version. If the brief owner confirms the portrait execution only needs the arrival and discovery, route three removes the additional setup, though the short version stops carrying the misunderstanding.

What limits does the piece place on the AICP reference and on what the treatment can conclude?

The AICP guide is US production-association guidance in question form; it sets no rate, establishes no entitlement or approval, and confirms nothing about feasibility for a particular setup. The treatment can name the creative dependency and route scheduling, cost, technical checks, and approvals as questions; it cannot determine affordability, schedulability, or permission. Two facts would change the recommendation—whether the short may say something different, and whether a portrait extraction holds enough of the far end—and neither has been tested.

Further reading

• Use an Animatic to Answer a Timing Question, Not Pretend the Commercial Is Finished — If the unresolved question is whether a compressed wait still reads, this explains when a timed rough assembly can settle a duration question drawings cannot.