How to Translate Subtitles Without Breaking the Timing

Translating a subtitle file is not the same job as translating a document. The words have to fit inside timing that was set for a different language, and the thing that breaks is almost never the translation itself. It is the fit.

Timing survives, reading speed does not

A subtitle translation should keep every timestamp exactly as it was. Cue one still starts and ends where cue one started and ended; only the text inside changes. That is what makes translation safe: the sync you already checked stays checked.

What does not survive is reading speed. Languages do not carry the same meaning in the same number of characters, and a translation commonly comes out longer than the English it came from. The cue duration has not changed, so a line that was comfortable at 16 characters per second can land well past 20 in the translated file. The timing is still correct and the captions are now too fast to read.

This is the check people skip. After translating, run the new file through the subtitle validator and look at reading speed and line length, not at sync. The reading speed guide covers what the numbers mean.

Translate from a reviewed source, not a raw draft

Errors multiply through translation. A name transcribed wrong in the source becomes a name translated wrong in every output language, and it is far harder to spot once it is in a language you may not read. Fix the source first: names, places, product terms, and anything spelled unusually.

Build a short list of terms that should not be translated at all. Brand names, product names, handles, and hashtags usually belong in the target file untouched, and a translation pass will happily convert them if nobody says otherwise.

What to check in the translated file

• Reading speed per cue, because the text got longer and the clock did not.
• Line length and line breaks. A break that fell at a natural phrase boundary in English rarely falls in the same place after translation.
• Names and brand terms, which should have come through unchanged.
• Numbers, dates and units, which are formatted differently across languages.
• The cue count. It should be identical to the source. A different count means cues were merged or split and the timing no longer matches.

Right to left languages need a second look

Urdu, Arabic, Persian and Hebrew read right to left, and a file that looks fine in a text editor can render badly in a player. Check the translated track in an actual player rather than trusting the file view, and save the file as UTF-8 so the script survives the trip at all. The subtitle encoding guide covers what goes wrong there.

Where to run the translation

Translation belongs in the same place as the review, so the timed text stays editable while you check the fit. Start from the video in the AutoCaption caption generator or go straight to the video subtitle translator, which keeps cue timing while the text changes and exports SRT, VTT, plain text, or a captioned video for each language. If a destination needs the other subtitle format afterwards, the SRT to VTT converter handles it.