Dubbing or subtitles? Settle it on one scene, not a whole season.

A translated season can still fail the viewer

A title is not localized because its dialogue has been translated. It is localized when the audience in that market can follow the story, the lesson or the argument in the way they actually watch. A dub replaces what the viewer hears. Subtitles leave the original performance audible and add something to read. Each spends the viewer's attention differently, and that difference, not habit or the size of the budget line, is what the choice should rest on.

The choice also hardens early. Once voices are cast for a series, changing format or cast midway is something the audience notices. A subtitled season that proves hard to watch on the target screen is not rescued by better translation. The time to be wrong is before the season is committed, on material small enough to redo.

Start with where the viewer's eyes have to be

Ask what carries the meaning on screen. When the picture does the work, as in a software walkthrough, a safety procedure, a cooking step or a fast action sequence, every second spent reading the bottom of the frame is a second not spent on the thing being shown. A dub or a voiceover frees the eyes. When the voice does the work, as in an interview, first-person testimony, a performance-led drama or comedy that lives on timing, subtitles keep the original delivery intact, and a replacement voice has to earn back what it removed.

Then look at density. Subtitles are a condensation, not a transcript: fast or overlapping dialogue has to be shortened to stay readable, and something is lost each time. A compliance explanation or a rapid comic exchange feels that loss far more than a slow, sparse scene.

Screen and setting count as well. A subtitle that sits comfortably on a television can crowd a phone or collide with on-screen graphics, and a clip likely to be watched with the sound off needs readable text whatever else it carries. And consider who is reading: an audience that reads slowly in the target script, or not yet at all, as with young children, needs to hear the language.

Then ask why they are watching

Purpose sharpens the answer. A viewer who watches to learn and then act, such as completing a training module or following an installation, is best judged by whether they can perform the step afterwards, and that tends to favour hearing the language while watching the hands. A viewer watching for entertainment wants immersion, which either format can deliver, depending on what the audience in that market is used to. A viewer watching to hear a real person, as in news, a documentary interview or a testimonial, may distrust a voice that pretends to be the speaker.

That last case is where a third option earns its place. Voiceover narration, one target-language voice laid over the lowered original, keeps the speaker audible without asking the viewer to read. It is neither a dub nor a subtitle, and a strict two-way choice would miss it.

What each format commits the production to

Subtitles need translation that fits time and space, timing against the picture, sensible line breaks, separate localization of any on-screen text, and a check on the actual delivery platform. A file with accurate words and poor timing is not a finished viewing experience. Their advantage is that a correction is a file edit.

A dub adds script adaptation, so that lines fit the length and, for on-camera speakers, the mouth movements of the original. Casting, direction, recording, mix and sync review follow. A late script change can mean studio time rather than a text fix, which matters for content that is revised often. A voice also carries accent and register in a way text does not, so a dub that sounds natural in one region can sound foreign in another; Spanish for Spain and for Latin America is the familiar case. For a series, voice continuity across episodes is part of the commitment.

Neither format settles accessibility. W3C accessibility guidance separates captions, which serve viewers who cannot hear the audio and include relevant sounds and speaker cues, from translated subtitles, which may carry dialogue only. A dub does not remove a captions requirement, and audio description is its own track. Where the platform, the market or an internal policy calls for them, list them as separate deliverables from the start.

Test one demanding scene before the season

Choose the scene most likely to break a format, not the easiest one: fast or overlapping speech, one moment where performance matters, one piece of on-screen text or a visual step the viewer must follow, and one culturally loaded line such as a joke or an idiom. Write down what a pass looks like before anyone watches. Useful criteria are comprehension, visual attention, whether the key moment lands, and whether the language sounds natural for that market.

Produce the scene in each candidate format and play it on the screen the audience will actually use, to viewers who resemble that audience. Bilingual reviewers are needed to diagnose problems, but they cannot stand in for the audience: they hear the source dialogue and fill gaps a target viewer cannot. Ask viewers to retell what happened or perform the step, rather than which version they liked. Record what failed and why.

The answer can differ by market: a dub where the audience watches on phones, subtitles where the original voice is the point, voiceover for the interviews, captions alongside whichever is chosen. What matters is that the reason is written into the brief before the whole season is produced. Once the format is fixed, the tracks still have to be scheduled against the release date, which the media localization launch window article takes up.

Where MoniSa fits in the decision

MoniSa's multimedia line covers subtitling, captioning including subtitles for the deaf and hard of hearing (SDH), dubbing, voiceover and audio description. Because the candidate formats sit in one scope, a test scene can be produced in each and judged against the same written criteria, and a season brief can record different answers for different markets without splitting the work into separate plans. Target-language linguists can then explain why a version failed, line by line, which is the part of the test a preference poll cannot supply.

Before the format goes into the season brief

Settle these on paper for each target market before production is costed.

• Where the viewer's attention must be during the scenes that carry the meaning
• Why this audience is watching: to learn, to be entertained or to hear a real person
• The screen and setting the audience will actually use, including muted viewing
• The test scene, chosen because it is hard, with pass criteria written before viewing
• The accessibility tracks required, listed separately from the language format
• The regional variety each dubbed or subtitled track must serve

Signs the format was chosen for the wrong reason

Any one of these means the choice has not yet been tested against the audience.

• The format was set from the budget line before anyone watched a scene
• The test scene is the calmest, slowest scene in the title
• Only bilingual reviewers have judged the versions
• Viewers were asked which version they preferred, not what they understood
• Translated subtitles are assumed to satisfy a captions requirement
• One dubbed track is expected to serve every region that speaks the language

What to send MoniSa to scope a test scene

Send enough for the test to be scoped against the real content rather than a generic recommendation.

• The candidate scene, or the title with the hardest scenes marked
• The target markets, the audience in each, and why that audience is watching
• The delivery platform and the main screen type
• The formats under consideration, including voiceover if it is in play
• Any captioning, SDH or audio description requirement already known