How to translate an SRT file (and keep the timing)

An SRT file looks intimidating until you open one. Then it looks like this:

12
00:01:04,200 --> 00:01:06,900
I've been waiting for you so long.

13
00:01:07,400 --> 00:01:08,850
Then why did you leave?

That's the whole format: a cue number, a start and end time, the text, a blank line, repeat. Which means translating an SRT file is conceptually simple — change the text, never touch the timestamps — and everything that goes wrong goes wrong by breaking that rule.

Why translations break timing

If you've ever run a subtitle file through a generic translator and gotten back something unwatchable, one of these usually happened:

The structure got mangled. Tools built for prose don't know that cue numbers and timestamps are sacred. They "translate" 00:01:04,200 into localized number formats, merge blocks, or drop the blank lines that separate cues. The file stops being valid SRT and your player ignores it or displays garbage.

Lines were translated in isolation. Subtitles split sentences across cues. "I never told you…" / "…what happened that night." translated cue-by-cue gives you two fragments that each read like a complete, wrong sentence. Names drift too: the same character comes back as three different spellings across an episode.

The text no longer fits the time. This is the subtle one. A cue that shows for 1.4 seconds can only hold so many characters before nobody can read it. Some languages run 30–40% longer than the source; a "correct" translation that overflows the cue is still an unwatchable one. Good subtitle translation re-breaks lines and respects reading speed; word-for-word translation doesn't.

The manual route

If it's a short clip and you speak both languages, honestly — a text editor works:

1. Make a copy of the file (always keep the original).
2. Open the copy in any plain-text editor (not Word).
3. Replace only the dialogue lines. Leave cue numbers, timestamps, and blank lines exactly as they are.
4. Save as UTF-8 with the .srt extension.

A free tool like Subtitle Edit or Aegisub makes this nicer — they show the video alongside the cues, so you catch overflow as you type. For anything longer than a scene, though, hand-translating hundreds of cues is a weekend, not a task.

The spreadsheet trick (and its ceiling)

A popular halfway approach: paste all the dialogue lines into a spreadsheet column, run them through a translator, paste back. It preserves the timestamps, which puts it ahead of most tools. But it inherits every line-in-isolation problem above — split sentences, drifting names, no reading-speed control — because each row is translated blind, with no idea what came before it.

What "keeping the timing" should actually mean

A translation worth watching keeps more than the timestamps:

• The original cue timing, untouched — your file stays in sync with the exact video it was made for.
• Context across cues — sentences that span cues are translated as sentences, and character names stay the same in cue 12 and cue 1,200.
• Reading speed and line breaks — text is re-wrapped to fit the screen and the seconds it's on screen, in the target language's own rhythm.

That last part is the difference between translated and watchable, and it's exactly the part that generic tools skip.

Langelic does the whole job in one pass: upload the SRT (ASS and SSA work too), pick your language, and download a file with the original timing intact and the text reflowed to read comfortably. If the file has sentence-splitting cues, recurring characters, or an episode's worth of running jokes, they're translated with the surrounding context in view.

Translate an SRT file now →