Dub or Subtitle? Explain the Localization Plan for a Finished TV Series

"Delivered dubbed and subtitled" is a sentence that shows up in a lot of finished-series pitches, and it is not a plan. It names two products and says nothing about the twenty seconds where a buyer's programmers will decide whether the show works in their language.

The answer up front: don't compare dubbing and subtitling in general. Pick one scene that cannot be judged by its words alone, describe what each route does to that scene beat by beat, then write down which of those effects you are choosing and which you are accepting. A plan made this way is specific enough to survive a buyer's questions, honest enough to survive delivery, and short enough to sit in a pitch.

Everything below works through one invented scene. The series, the characters and the exchange do not exist; they are a teaching object. I have not translated, recorded, timed or screened any version of it, and the specialist judgments it needs — a translator's, a dub director's, a performer's — are ones you will have to go and get. The point is to show which questions to bring.

Define what is being localized before you argue about how

A language edition is made from a finished program. That is a different transaction from licensing the format so another company can remake the show in its own market, and it is different again from translating the pitch deck. The first produces a new production you do not control. The second produces a document. Neither one is a language edition of the episodes you are selling, and buyers notice when a seller blurs them together, because the rights, the costs and the people doing the work are entirely different.

So start with the inventory, in plain terms.

Picture. The finished masters. Note whether any text is burned into the image — a shop sign, a phone screen, a title card, a handwritten label. Text inside the picture has to be dealt with by both routes, and by neither automatically.

Dialogue. The recorded performances, and the writing underneath them. A translated subtitle file and a dubbing script are two different deliverables made from the same source by different processes. One is a timed text track; the other is a script that performers will speak.

Sound. The final mix, and whether the dialogue can be separated from everything else. Dubbing generally needs a version of the soundtrack with the voices removed and the music and effects intact — often called an M&E — because the new performance has to be laid onto the same music and the same room. If the only thing that exists is the finished mixed audio, with the dialogue baked into the music, that is a production dependency to solve, not a detail to promise past.

Then label the state of each item in three columns: what exists, what is commissioned, and what is merely proposed. In most pitches at this stage, the honest answer is that the masters exist, the translation does not, the dub script does not, no performer has been cast and no timing has been checked. Writing that down is not weakness. It is the difference between a plan and a wish, and it tells the buyer exactly what their money is buying.

Find the scene that a word count cannot settle

Here is the invented scene. It runs nineteen seconds, in episode four of a six-part drama.

At 0:00, Ruth sits alone at a folding table in a laundromat at night, holding a small folded note. The audience sees the back of it. She reads it once, with no change of expression. Room tone; a kettle somewhere off screen.

At 0:11, Theo walks into frame, off to her left.

At 0:12, he says one line. Roughly: You always read the good ones twice. It is a tease, and in this invented source version it is spoken without naming who is being teased. Whether it is aimed at Ruth's sentimentality, or at Theo's own exposure, is left open by the words.

At 0:15, the line ends. Ruth looks up and holds Theo's gaze. Nothing is said.

At 0:16, a single sustained music cue enters underneath.

At 0:17, her eyes drop to the folded note in her hand, then rise back to Theo. The shot cuts at 0:19.

That last look is the whole scene. The audience never learns what the note says — they only learn whose business it is. The note concerns Theo. And the moment that lands, the joke turns around: the tease is now about him, and the audience is left suspecting he would rather she had not read it at all.

Three things have to arrive, in this order. The line has to register as a tease. Its target has to stay unsettled through 0:15. And the target has to be decided by the picture and the music at 0:17, not by anything spoken. The line is short, the drop of the eyes is one and a half seconds, and everything the scene means sits on a look.

That is the kind of moment worth building a localization plan around, because it refuses to be measured in words. If your plan can describe what should happen here — before anyone has translated anything — you have something a buyer can actually evaluate.

Compare what each route does to the scene

Subtitling keeps the original performances. The viewer hears Ruth and Theo as they were recorded, and reads a translated line while watching their faces. The translated subtitle for Theo's line will typically be displayed a little wider than the speech itself, so that readers have a moment to find it — which means the text is plausibly still on screen at 0:15, when the hold begins, and possibly into the first half-second of the glance.

Dubbing replaces the voices. The viewer understands the line without reading, and their eyes stay with the faces. In exchange, the original performances are gone, and the line now exists as a new series of choices: how Theo's replacement sounds when he says it, how much weight he puts on good, whether the tease is delivered as settled or unsettled. That delivery can resolve the target early just as easily as a subtitle can.

Lay the routes over the beats and the trade becomes visible.

Two things follow. First, neither route is automatically better here; the failure mode is nearly identical — something resolves the tease before the look does. Second, the routes differ in what they cost when they succeed. The subtitled edition spends the viewer's attention and keeps the original voices. The dubbed edition protects the viewer's attention and spends the original performances, which for a six-part drama with a lead character means a continuity obligation across every episode, not just this scene.

Notice also what the plan should not claim. Nobody can tell you from a desk whether the target language can leave the tease unattributed the way the invented source line does. That is a question for a translator looking at the real script, and the answer may be that the target language forces a choice — in which case the question becomes where the choice ought to land, and whether the line can be rewritten to keep the look doing the revealing. A machine translation can help you see whether a line is in trouble. It is not a translated script and it is certainly not a dub, and it should never appear in a pitch as either.

Turn the preference into a bounded proposal

Here is how the two proposals would read before a translator and a dub director have been through them. Treat them as drafts awaiting that review, because that is what they are.

Proposal A — subtitled edition of the finished masters. Priority: the tease stays unattributed until 0:17, so the reveal belongs to the picture and the music cue. To commission: a translation briefed on that constraint, with the scene's ambiguity written into it as a requirement rather than a preference; a timing pass that checks the display window of Theo's line against the hold and the music cue at 0:16; a decision from the translator about what to do if the target language must name a target, including whether the line can be recast to avoid it. Accepted compromise: some viewers will be reading during the hold, and the reveal will land a half-beat later for them than for anyone else.

Proposal B — dubbed edition. Priority: the same. The look must still land. To commission: a dub script that preserves the withheld quality of the line; a performer and a director briefed not to settle the tease on the word good; a timing review against the fixed picture, since the glance cannot be moved; and a decision about the rest of the series, because a recast voice is a commitment across six episodes. Accepted compromise: the original performance choices — the ones that made the scene worth protecting — are no longer available to the viewer.

Neither of these is finished work, and neither can be until someone is paid to do it. What they do is make the choice visible.

Now the pitch sentence. Something in this shape:

The six finished episodes of The Quiet Floor will be prepared for [market] in [language]. One priority governs the language edition: in episode four, scene eleven, the tease at 0:12 must remain unattributed until the look at 0:17. We propose a subtitled edition, translated and timed under that constraint, because the reveal depends on the recorded performances staying in place. A dubbed edition is possible and we would like to discuss it; it would require a music-and-effects element, a dub script, casting and a timing review, none of which is in place today.

That paragraph is a plan. It names a scene, states a priority, proposes one route, offers the other, and separates what exists from what would need commissioning.

Keep three other distinctions clean while you write it. Translated subtitles are not the same deliverable as captions for access: captions assume the viewer cannot hear and carry speaker labels and sound information that a translation track does not. Audio description is a third job again, narrated for viewers who cannot see the image. They may travel in the same package, and delivering one is not evidence that the others are done. And none of this settles rights, performer agreements, commissioning budget or the recipient's delivery specifications, which are theirs to state and yours to verify separately. A creative recommendation is not a compliance check.

What the scene buys you

The scene gives you a route, and a trade-off, and a list of work. For this one, a subtitled edition is the better first proposal, because the thing worth protecting is a performance and the thing doing the revealing is already in the picture. The dubbing conversation stays open — and stays honest, as long as the music-and-effects element, the script, the cast and the timing pass are described as outstanding.

What you have not done, and should not pretend to have done, is translate the line, record the dub, time the subtitle, or ask a single viewer what they saw. Those are the next pieces of work. The value of the plan is that everyone reading it now knows exactly which pieces they are.

Frequently asked questions

What turns a finished-series localization pitch from a label into a plan?

Pick one scene that cannot be judged by its words alone, describe what subtitling and dubbing each do to that scene beat by beat, then state which effects you are choosing and which you are accepting. Also inventory what exists, what is commissioned, and what is merely proposed, rather than saying the show will be delivered dubbed and subtitled.

Why build the plan around one difficult scene instead of comparing dubbing and subtitling in general?

A general comparison misses where a scene's meaning sits. In the invented scene, Theo's line at 0:12 must register as a tease, stay unattributed through 0:15, and let the look at 0:17 decide the target. Subtitles can still be on screen during the hold, while a dubbed delivery can settle the target early. The scene makes those risks concrete.

What must be separated before arguing about dub or subtitle?

A language edition is different from licensing a format for remake and from translating a pitch deck. Inventory picture, dialogue, and sound, and label each as existing, commissioned, or proposed. Burned-in text affects both routes, and dubbing generally needs an M&E with dialogue removed if the finished mix will not separate it.

What does each route cost in the invented scene?

Subtitling keeps the original performances but spends viewer attention and may leave translated text on screen into the hold. Dubbing protects attention by removing the reading but spends the original performances, and delivery can resolve the tease early; for a six-part drama with a lead, that also creates continuity obligations across episodes. Neither route is automatically better.

What can the plan not claim before a translator or dub team works on it?

It cannot claim the target language will leave the tease unattributed; that is for a translator looking at the real script, and the language may force a choice, in which case the question becomes where it should land or whether the line can be rewritten. Machine translation is not a translated script or a dub. For dubbing, the M&E, dub script, casting, timing review, and continuity decision are outstanding. Translated subtitles are also not the same deliverable as access captions or audio description.

Further reading

• A Film’s Joke Depends on Its Language. Translate the Words—or Rebuild the Joke? — Useful when the line's effect depends on wording rather than plot, and you must choose between subtitling the joke and rebuilding the scene around it.
• Read a TV Availability Sheet Before Copying Its Claims Into a Sales Pitch — Use this before repeating a language claim from an availability record, since it distinguishes offered, intended and delivered languages and flags unspecified asset readiness.