Explain the Film's Subtitles as a Creative Choice

"About a third of the dialogue is in [A] and it stays unsubtitled" is a sentence a pitch can carry for months without anyone deciding what a viewer actually knows. It is accurate. It is also the same sentence whether the scene plays as a mystery, a joke, an act of cruelty, or a mistake.

A reader of the pitch — a funder, a co-producer, a sales agent — is trying to imagine sitting in the dark. What they need from you is a scene-level answer to a simple question: at this moment, what do I know?

For one scene, three things supply that answer:

• Viewer knowledge at each beat. Who understands what, and when understanding arrives.
• Where the words sit and when they appear. The cue's placement against the image and its in and out moments.
• What the omission is for. The job the missing translation does that the scene needs done.

Here is a scene built to practice all three. It is invented, and the two languages are labelled A and B rather than written, because inventing plausible foreign dialogue for an example would just add a fake nobody can check.

Follow the exchange without knowing either language

The setup: Nour opens her door. Ilya is on the step with a cloth bag and speaks A. Mara, behind Nour, speaks A and B. Nour speaks B only. Inside the bag is a bowl that Ilya has had repaired, and at the door he names it in A — before Nour has seen it. A door chime starts the scene.

The lines below are meaning briefs: what each line has to accomplish. They are not dialogue.

• Beat 1. Door chime. No speech.
• Beat 2. Ilya, in A: a greeting, and an apology for arriving early.
• Beat 3. Ilya lifts the bag and, in A, names what is in it and who mended it.
• Beat 4. Mara's face changes.
• Beat 5. Nour, in B: What did he say?
• Beat 6. Mara, in B: Come in. It's cold.
• Beat 7. Later, Nour opens the bag.

The film's strategy in this version: the track carries B, and carries A only where the film chooses. So B is always readable, and A is readable at beat 2 and silent at beat 3.

Now take the diagnostic position of a viewer who knows neither language. Not because that viewer is representative — they are not — but because they set the floor of what the scene delivers without any linguistic help. Everything above the floor is real information too, and worth logging separately.

Read down the third column and you have the pitch's real claim. Read down the second and you have the drama. The gap between them is the scene.

Two things this exposes that a language list never will. First, the viewer is being asked to hold two unfamiliar languages apart, and the film has to help — through the sound of them, through performance, or through a subtitle convention such as italicising one and not the other. If a viewer cannot tell A from B, a selective strategy has no shape. Say how you expect them to tell. Second, every character is tracking something different from every other character, and the pitch only earns its confidence if it has actually written those differences down.

Compare full and selective translation

Now put a fully subtitled version beside the selective one. Keep the shots, the performances and the cut identical. Change one cue, and watch the same footage turn into a different scene.

In the full version, beat 3 carries a subtitle. The viewer knows about the bowl at the door. Beats 4 through 6 change meaning immediately: Mara's reaction becomes confirmation instead of a puzzle; Nour's What did he say? becomes a question the viewer can answer and she cannot; Mara's Come in. It's cold becomes a visible act of withholding rather than just a deflection. Beat 7 stops being a discovery and becomes a payoff.

In the selective version, the viewer spends beats 4 through 6 in Nour's position, and the bag opens for both of them at once.

Neither is better in the abstract, and a pitch that implies otherwise is dodging. What the pitch has to say is which one, and why this scene needs it. If the scene's job is to put the audience beside Nour, the cue at beat 3 comes out. If its job is to let the audience watch Nour not know, the cue stays.

Three warnings about the reasoning that gets written here.

A missing subtitle does not read the same as noise. A viewer can see that a mouth is moving. If the track has been reading A up to this point and then goes quiet on the one line that matters, some viewers will read that silence as a deliberate beat and some will read it as a subtitle house's error. The difference is usually made by what else happens — Mara's face, a held shot, a cut that isn't there. If you are relying on the absence, say what is carrying it.

An omission needs a purpose beyond atmosphere. The test: substitute a language unrelated to anything in the film's world, or an invented one, and see whether the scene still works. If it does, the specific language is decoration and the speaker has been reduced to texture. Sometimes that is a legitimate choice, but it should be a named one — the viewer shares Nour's exclusion here — rather than a mood the pitch hopes will attach itself.

Translate-later is a third option, and it costs something. You can withhold the cue at beat 3 and place it over the open bag at beat 7, so the words arrive with the object. That can be effective. It also puts a cue far from the mouth that spoke it, which some viewers will read as a sync problem, and it leaves the line floating without an obvious speaker unless Ilya is still in frame. Test it against the shot list before you promise it.

Once you have chosen, note the obvious consequence honestly: a viewer who understands A gets the contents of the bag at beat 3 regardless of what your subtitle track does. Selective translation distributes information unevenly, and that is a thing to state, not a thing to hide.

Show the words sharing the frame

Subtitles go in the lower part of the frame, which is also where hands, bags and objects tend to be. Your bag is being lifted into exactly that region.

So annotate the shot, not just the line. At beat 3 in the full version — the naming line, and the only cue that differs between the two versions — ask where the cue sits relative to the bag. If it lands over the bag, the viewer reads the name of what is inside while the bag's shape competes with the word for the same patch of attention — and for the seconds the cue is up, the reader is reading rather than looking. If it sits over Mara's face, you lose the moment when she understands before Nour does. Giving it back means moving the cue higher, which can push it onto faces, or placing it at the top of frame, which is a less expected position and can read as a mistake to viewers who have been trained by everything else they have watched.

Timing does the same kind of work with no repositioning at all. A cue that arrives before Ilya lifts the bag tells the viewer first and lets them look second. A cue that arrives after the lift lets the viewer see the shape of the thing, wonder, and then read. Those are two different orders of information, and you can state either as an intention today.

Whatever you propose, compare the alternatives on the same exchange. If you also cut to a close-up on the bag, or extend the shot by a second, you have changed the scene as well as the subtitle, and you cannot tell which change produced the effect you wanted. Hold the cut still. Move one thing.

Keep translation and access requirements distinguishable

Three jobs get done on screen here, and in practice they are often muddled together because they may end up in the same file, rendered by the same person, at the same time.

Subtitling translates speech the audience cannot otherwise follow. Creative on-screen text is writing the film chose to put in the image — a sign, a letter, a title that belongs to the world. Captions carry dialogue and meaningful non-speech sound for viewers who are deaf or hard of hearing: the door chime, the latch, who is speaking.

The vocabulary overlaps, and so do the assets. The jobs do not. Nour standing at the door hearing a chime is a moment that a subtitle track cannot deliver to anyone, because there is nothing to translate. If the chime matters — and in this scene it is the one piece of information every viewer receives on identical terms, in any language — then that fact belongs to your sound design and to the caption track, not to your subtitle strategy. A pitch sentence that says "the door sound is important" without saying which of those two it is pointing at has said nothing yet.

There is a harder version of this in the selective strategy, and it deserves to be named rather than smoothed over. A caption track exists to convey dialogue. If it does its job at beat 3, the caption user learns what is in the bag at the moment the subtitle audience does not — which means the carefully built exclusion at beats 4 through 6 does not apply to them. That may be acceptable. It may suggest restructuring the scene so the withheld line carries less weight. It may have an answer I should not invent here. What I can say is that this is not a subtitle decision, and it is not a compliance question for a pitch to settle either. It belongs to the filmmaker and to whoever is building the access track, and the useful thing a pitch can do is show that you know the two are related.

Which captions are required, and in what form, depends on where the film is shown and who is handling delivery. That is a question for those people. Do not let a creative rationale quietly stand in for an answer you have not got.

State the proposed strategy and its unresolved checks

Here is the whole method compressed into a pitch paragraph for this scene. It follows the selective route.

Viewers who do not speak A follow Nour's knowledge, not Ilya's. Ilya's greeting is translated at the door; the sentence where he names what is in the bag is not. The audience sees the bag, sees Mara understand something, and sees Mara answer Nour with Come in. The contents stay closed until Nour opens them. Viewers who do speak A have known since the doorstep, and the scene does not pretend otherwise.

That is a paragraph someone can check. It names the viewer's position at a beat, the placement decision, and the thing being withheld. It does not promise that everyone will feel the same thing, because everyone will not.

Then mark what is not settled, and be specific about who settles it:

• Meaning. Whether Ilya's line says what the brief claims, and whether the A lines as performed carry the register the scene assumes. This needs a qualified speaker of A — a translator, or the writer if A is theirs — reading the lines and the brief side by side.
• The silence at beat 3. Whether it reads as a deliberate hold or a dropped cue. That depends on the performance and the shot, neither of which exists yet.
• The caption track. What a dialogue-carrying caption does with an untranslated line, which is a conversation with an access practitioner, not a decision to make alone.
• The label problem. How a viewer with no A and no B is meant to tell the two languages apart on first hearing.

What exists now is a paper intention. No line has been translated, no cue has been timed, and no access review has happened. That is the correct state for a pitch — but the pitch should say so rather than let a confident paragraph imply finished work.

End on the scene, not on the promise

The honest close is one annotation and its open questions. Something like: Beat 3: Ilya names the bowl in A. No subtitle. Mara's reaction carries the weight. Viewer and Nour learn together at the bag. Check the A line with a speaker; check what the caption track does here; check whether the hold survives the performance.

That is a defensible claim about a scene, with its limits attached, and it is worth more than any assurance that an audience will feel the right thing. You are not promising a shared experience. You are telling a reader exactly where the viewer stands at each moment, and letting them decide whether that is the film they want to make.

Frequently asked questions

What does a pitch need to say about subtitles beyond how much dialogue is untranslated?

It needs a scene-level answer to what the viewer knows at each moment. For one scene, three things supply that: viewer knowledge at each beat, where the words sit and when they appear, and what the omission is for: the job the missing translation does that the scene needs done.

How does full translation change the example scene compared with selective translation?

In the full version, beat 3 carries a subtitle, so the viewer knows about the bowl at the door. Mara’s reaction becomes confirmation instead of a puzzle; Nour’s question becomes one the viewer can answer and she cannot; Mara’s invitation becomes visible withholding; and beat 7 becomes a payoff. In the selective version, the viewer spends beats 4 through 6 in Nour’s position and learns at the bag with her. Neither is better in the abstract; the pitch must say which one and why the scene needs it.

What are the risks of withholding a subtitle on a line that matters?

A missing subtitle does not read the same as noise. If the track has been reading A and then goes quiet on the one line that matters, some viewers may read it as a deliberate beat and others as a subtitle error. The difference is usually made by what else happens, such as Mara’s face, a held shot, or a cut that is not there. If relying on the absence, say what is carrying it. An omission also needs a purpose beyond atmosphere.

How do subtitle placement and timing affect the scene?

Subtitles go in the lower frame, where hands, bags, and objects often are. At beat 3, if the cue lands over the bag, the viewer reads the name of what is inside while the bag’s shape competes for attention; if it sits over Mara’s face, you lose the moment she understands before Nour does; moving it higher can push it onto faces, and top-of-frame placement may read as a mistake. Timing also matters: a cue before the bag is lifted tells the viewer first and lets them look second; a cue after the lift lets them see the shape, wonder, and then read.

How are subtitling, creative on-screen text, and captions different?

Subtitling translates speech the audience cannot otherwise follow. Creative on-screen text is writing the film chose to put in the image, such as a sign, a letter, or a title that belongs to the world. Captions carry dialogue and meaningful non-speech sound for viewers who are deaf or hard of hearing, such as the door chime, the latch, or who is speaking. The jobs can end up in the same file but are not the same job; the door chime belongs to sound design and the caption track, not to the subtitle strategy.

Further reading

• Add Captions to a Film Pitch Clip and Deliver the Right Version — When the scene's caption track moves from note to actual clip, this covers cue timing against one cut, transcript accuracy, separate file versus burned-in text, and testing the receiving route.
• Build Audience Access Into the Film Proposal, Not Just the Final Export — Useful when the caption strategy needs a place in the proposal; it separates captions from translation subtitles and keeps intended uncertainty from being explained away by access materials.