Subtitle Translation Rules: Character Limits and Reading Speed

Seen from outside, subtitle translation looks like the easiest kind of translation: the sentences are short, the language is spoken language. In practice it is the most constrained kind there is. Subtitle translation rules settle in advance how long the text may be, how long it will stay on screen and where it will break.

Contents

These constraints are not arbitrary. The viewer has to watch the picture and read the text at the same time; reading time is time stolen from the picture. The rules were developed to protect that balance.

Below we take up what the rules are, how they are applied and which decisions come to the fore in Turkish subtitling.

Why are subtitle translation rules more than translation?

In document translation the single criterion is that the meaning is carried across correctly. In subtitling three criteria have to be met at the same time: the meaning must be right, the text must fit the duration, and the line break must be readable.

These three clash constantly. The correct translation does not fit the duration; the translation that fits the duration thins the meaning; the line break that preserves the meaning spoils readability. The subtitle translator’s job is to keep striking a balance between the three.

Then there is a fourth layer: the picture. Information already visible on screen is not repeated in the subtitle. If the speaker is showing a document, there may be no need to say what the document is. This is the most fundamental area of decision that separates subtitling from text translation.

Subtitle translation is, in the end, not a transfer but a rewriting. That is also why it is priced by the minute rather than by the word: the workload is proportional to the duration, not to the volume of text.

Character limits and line layout

The number of lines in each subtitle block and the number of characters per line are limited. The limits vary from broadcaster to broadcaster; in common practice a subtitle consists of at most two lines.

Different criteria are used across the industry for the character limit per line. In the timed text style guide Netflix publishes, for example, the general rule is 42 characters per line; at other broadcasters that figure can vary between 37 and 40. The limit is set according to screen width and font size.

The practical consequence of the limit is this: if the source sentence is long, it is either compressed or split across two subtitles. The decision to split is not an easy one; a subtitle broken in the middle of a sentence makes the viewer grasp the meaning late.

An automatic check is therefore run before delivery. Blocks that exceed the character limit, subtitles running to more than two lines and empty blocks are listed. Done by hand, this check lets things slip; the quality control modules of subtitling software complete the job in seconds.

Reading speed and duration limits

Reading speed is the most decisive constraint in subtitling. It expresses how many characters can comfortably be read in a second and is measured in characters per second (CPS).

As a widely used criterion, the Netflix style guide applies a limit of 17 characters per second on adult content and 13 characters per second on children’s programmes. The same guide also defines a lower and an upper limit for how long a subtitle stays on screen; a subtitle that stays too briefly cannot be read, while one that stays too long makes the viewer read the text over again.

These limits determine translation decisions directly. If the speaker talks fast, the text has to be shortened; once the limit is exceeded, the subtitle counts as technically faulty and is not accepted.

Because different broadcasters have different limits, it has to be settled before the project begins which style guide the work will follow. If the guide changes later on, the whole file may have to be reviewed again.

Text compression: what do you leave out?

Compression is the core skill of subtitle translation. The aim is not to shorten the text but to shorten it while keeping the meaning intact.

The order of removal usually runs like this: first repetitions and self-corrections, then filler expressions (“yani”, “işte”, “şey”), next forms of address and speakers’ names, and last of all information that is clear from the picture. Content that carries meaning is left to the very end and, where possible, not removed at all.

1. Remove repeated words and self-corrections.
2. Drop filler expressions and sounds of hesitation.
3. Take out forms of address and calls by name where no meaning is lost.
4. Remove from the text the information plainly visible in the picture.
5. Replace long noun phrases with shorter equivalents.
6. If need be, split the sentence and spread it across two subtitles.

The limit of compression is style. The character’s manner of speech, the level of formality and the emotional tone have to be preserved; an over-compressed subtitle is accurate but lifeless.

The rate of compression varies with the content. In productions containing natural speech, such as interviews and documentaries, there is plenty of material that can come out; in scripted dialogue, where every line has been composed, the room is far narrower. In training videos containing technical exposition, compression is the riskiest of all, because every detail removed is a loss of information.

Time code, cuts and line breaks

The start and end times of a subtitle have to fit the flow of the picture. The basic rule is that the subtitle begins together with the speech and does not disappear before the speech ends.

The second rule concerns scene cuts. A subtitle that carries on past the moment of a cut makes the viewer uncomfortable; wherever possible, subtitles are aligned to the cuts. This is not merely a technical preference but a matter of readability.

The line break, for its part, is a grammatical decision. It is made so as not to separate units of meaning: separating an adjective from its noun, a noun from its suffix or an auxiliary verb from the main verb slows reading down. In Turkish this rule matters especially, because suffixes come at the end of the word and a break delays the meaning.

Finally, the number of subtitles is a criterion too. The same content can be given in fewer and longer subtitles; this lightens the reading load but makes synchronisation harder. The decision is taken according to the pace of the content.

The production flow: which step sits with whom?

A subtitling project consists not of a single job but of steps that follow one another. When you compare quotes you need to see which steps are included; that is generally where the differences lie.

The first step is transcription: putting the speech into writing. If you already have a subtitle file in the source language, this step is not needed. The second step is time coding, the job of establishing the start and end moments of the subtitles, and it calls for a separate expertise.

The third step is translation. The fourth is revision, carried out for language and for synchronisation alike. The fifth step is the technical check: character limit, reading speed, empty block and overlap controls are run automatically. The sixth step is delivery in the target format and, where required, burning the subtitles into the picture.

Not all six of these steps are needed on every project. If your source subtitle is time coded, the first two steps fall away and the cost drops markedly. That is why stating exactly what material you have when you ask for a quote is the quickest way to get the right price.

File formats and delivery

The subtitle file format is chosen according to the platform it will be used on. A file delivered in the wrong format cannot be used even if it has been translated correctly.

Format	Typical use	Formatting support
SRT	General purpose, players	Limited
WebVTT	Web players, HTML5	Positioning and styling
TTML / EBU-TT	Broadcasting and streaming platforms	Broad
ASS / SSA	Content that needs special typography	Very broad

The second critical point at delivery is character encoding. If subtitle files are not saved as UTF-8, Turkish letters appear corrupted in the player. This is one of the most frequently reported technical problems, and it goes unnoticed until the file is opened.

The third is the frame rate. Time codes are produced against a particular frame rate; matched with a picture at a different frame rate, the subtitle drifts further and further out. The frame rate of the source video has to be confirmed before delivery.

The Türkiye context

The most decisive linguistic feature in Turkish subtitling is that the sentence is built differently from the source languages. The verb, which comes early in English, shifts to the end of the sentence in Turkish. In subtitles split over two lines this leaves the meaning incomplete until the second line, and the viewer grasps the sentence late. Experienced subtitle translators therefore rebuild the sentence so that the first line delivers the core of the meaning, rather than carrying it across word for word.

The second feature is length. Thanks to the agglutinative structure, some expressions grow shorter; chains of suffixes and compound noun phrases, on the other hand, can stretch the text out. This fluctuation makes it necessary to check the character limit line by line in Turkish; a sentence that fits in the source language may not fit in Turkish.

The third is viewer habit. Although the tradition of dubbing is strong in cinema and television in Türkiye, watching with subtitles has become widespread on online platforms. This has raised expectations of subtitle quality: spelling mistakes and drifts in synchronisation are reported by viewers very quickly.

What this means for you

If you are buying subtitles for video content, what determines the quote and the quality of delivery is not the word count but the information below.

1. State from the outset which style guide the work will follow; if there is none, ask the provider to recommend one.
2. If you have a time coded source subtitle, share it; it brings down both the schedule and the cost markedly.
3. Give notice of the delivery format and the frame rate in writing.
4. Prepare a short list of proper names and terms; consistency of spelling comes from there.
5. Plan a round of synchronisation checking after delivery; most corrections come out at this step.

Once these five pieces of information are ready, subtitling projects become predictable. Share your video duration and your target platform and we can work out the scope: request a quote or look through our service scopes.

Frequently asked questions

Why is subtitle translation not priced per word?

Because the workload is proportional not to the word count but to the duration and the number of subtitles. Adjusting time codes, compression and synchronisation checking all demand labour independent of text volume. Pricing per minute is common for that reason.

What happens if I do not supply a time coded file?

The time codes have to be produced from scratch; this is a separate item and it lengthens the schedule. If you have a time coded source subtitle, the translation process speeds up markedly.

Is everything that is said translated in a subtitle?

No. Because of the reading speed limit, the text is compressed. Repetitions, filler expressions and information clear from the picture come out first; content that carries meaning is preserved.

What is the most common mistake in Turkish subtitling?

Trying to carry the source sentence across as it stands. Because Turkish sentence structure is different, a word-for-word transfer both runs long and spoils the line break. The right approach is to keep the meaning and rebuild the expression.

References

• Pietrzak, P. & Kornacki, M. (2021). Using CAT Tools in Freelance Translation. Routledge, pp. 17–18.
• Díaz Cintas, J. & Remael, A. (2021). Subtitling: Concepts and Practices. Routledge, pp. 85–120.
• Netflix Partner Help Center, Timed Text Style Guide: General Requirements – character and reading speed limits.
• W3C, WebVTT: The Web Video Text Tracks Format.