Why Translated Subtitles Break Timing: Text Expansion Explained

Take an English subtitle file with good timing, translate it into German and play it back. Nothing has moved, yet the subtitles suddenly feel rushed. The timestamps are the same; the text isn’t. A translation rarely needs the same number of characters to say the same thing, and in subtitles every character has to be read within a fixed number of seconds. This is called text expansion. It is predictable, and once you know where it shows up, you can fix it.

How much longer translations get

Localization teams have been measuring this for decades. It is the reason buttons overflow in translated apps, and the W3C keeps a handy reference on text size in translation. Here are rough figures for translations from English. Treat them as a guide: the real number depends on the text.

The table actually describes two separate problems. The first is expansion: the same sentence needs more characters. The second is word length. German and Finnish pack a lot of meaning into single long words that can’t be split across lines, so even a little expansion produces awkward lines. Chinese, Japanese and Korean turn this around. They use far fewer characters, but reading-speed limits for those scripts are set per character, so a shorter line isn’t necessarily quicker to read.

Why subtitles suffer most

Each subtitle cue is a deal between text and time: this many characters, readable in this many seconds. The usual limits of about 42 characters a line and 15–17 characters per second exist to keep that deal fair, as we explained in subtitle line length rules. Translation changes the text and leaves the time exactly where it was.

• A 38-character English line that read comfortably at 15 cps becomes a 50-character German line at over 20 cps. Very few people can read that fast.

• A line break that sat neatly between two English phrases now lands in the middle of a German compound. Nobody wants a line to end on “Geschwindigkeits-” with “begrenzung” waiting on the next one.

• Cues that fit on two lines in English now need three, which most style guides don’t allow, or they spread wider than a player can comfortably show.

How subtitlers fix it

1. Condense the translation. This always comes first. Subtitles were never meant to be a word-for-word record of what was said. Cutting filler and choosing shorter words while keeping the meaning is the core of the job, and a fluent, slightly shorter line beats a faithful one that nobody can finish reading.

2. Re-break the lines for the target language. Break points that worked for English phrases don’t work for German compounds or Russian case endings. Each language needs its own line breaks.

3. Re-cue where there’s time. Merge two short cues that carry one sentence, or let a cue run on into the pause after it. The video stays as it is; only the time slots for the text change.

4. Only then look at the video. If a market is worth the effort, slower edits or a dub can solve what text alone can’t. We weigh that up in subtitles vs dubbing.

A workflow that avoids most of the trouble

Text expansion is easy to handle when your process expects it.

• Translate the subtitle file, not plain text. A subtitle translator keeps cues and timestamps intact, so your review is about length rather than repairs. The mechanics are in how to translate SRT and keep the timing.

• Review the crowded lines in each language. Look for the densest cues, because that’s where the expansion ended up. A characters-per-second counter makes them easy to spot.

• Treat the source as the master copy. Fix mistakes in the original and translate again. Language files patched by hand drift apart, and every later edit then has to be made several times over.

• Write source text that translates well if the script is yours. Short sentences survive translation better. Localization style guides have told app developers the same thing for years, and public subtitle standards such as the BBC subtitle guidelines rest on the same reading-speed logic.

Frequently asked questions

Why do my German subtitles feel fast when the timing is identical?

Reading speed is measured in characters per second, and the German version has more characters. Same timing plus more text means faster reading. Condense the lines or break them differently.

Should I change timestamps for each language?

Rarely, and only as a last step. Condensing and re-breaking fix most cues without touching the timing. Re-cue only where a pause gives you spare seconds.

Is text expansion a machine-translation problem?

No. Human translators get the same expansion, because it comes from the languages, not from whoever translates. Machine translation just makes it easier to publish the crowded version without noticing.

Which languages need the most review time?

German, for length and compound words. Chinese, Japanese and Korean, for their own reading-speed rules. Arabic and Hebrew, for right-to-left layout. Expect your first translation into each language family to take longer, because you will be learning as you go.