How to Translate an SRT File to English: Three Paths Compared
You have an SRT file and want English subtitles. It looks like a single task — "translate" — but it is actually three problems stacked together: translate the text, keep the timeline intact, and assemble the result back into a valid file. What separates the three paths is who solves each of these three problems.
First, Look at What Is Inside an SRT
SRT is plain text with a structure so simple there is almost no room for error: a number, a timeline, text, with a blank line between cues.
1
00:00:01,000 --> 00:00:02,500
Hoy vamos a hablar de tres cosas
During translation only the text changes — the numbers and the timeline must be preserved exactly. Get the timing wrong by one second and the subtitles fall out of sync with the speech, which is hard to catch after the fact. Any approach that does not understand this structure puts the risk of reassembly on you.
Path 1: Human Translation
Fits: branded content, legal and medical material, independent film and documentary distribution. Consistent terminology and controlled tone are things machines still cannot fully deliver.
Cost: human subtitling is usually priced per minute of finished video, and the price stacks up as language pairs multiply; there is a delivery lead time, and small orders often come with a minimum charge — a 3-minute job may cost the same as a 30-minute one.
Hidden cost: without a glossary and background material, translators can only guess, and the back-and-forth review takes longer than expected — none of which appears on the quote.
Path 2: Pasting Cues into a Generic Translation Tool
Fits: one-off situations with only a handful of subtitle lines.
Problem: generic translation tools do not understand SRT structure. What you paste in is text with "number + timeline + text" mixed together, and what comes back is one continuous block that you have to re-split, re-time and reassemble into valid cues. Once the cue count grows, misalignment is almost guaranteed — and it shows up as a global shift or a few cues swapping places, which is very hard to check by eye.
Path 3: A Dedicated Online Subtitle Tool (Free, Recommended)
A dedicated tool first parses the structure, pulls the numbers and timeline out separately, sends only the text for translation, then assembles a new file on the original timeline. Of the three problems, the tool covers the last two; you only need to judge the translation quality.
Take SmileSub's SRT translator as an example — no login, no software install:
- Upload the SRT file
- Pick the source and target languages — 20 languages are supported for mutual translation, including Chinese, English, Japanese and Korean
- Wait for parsing and translation to finish
- Download the translated SRT, with the timeline and cue count unchanged
Free quota: single files up to 1 MB and up to 500 subtitle cues, with 3 uses per tool per IP per day. For long videos, split the content into sections and process them one at a time.
How to Choose Among the Three
| Your situation | Recommendation | Main cost |
|---|---|---|
| Only a few subtitle lines | Generic translation tool, handled manually | Your time |
| Professional, legal or medical content requiring signed-off delivery | Human translation | Per-minute pricing + lead time |
| Voiceover / tutorials / batch international distribution | Online tool | Almost zero (mind the free quota) |
The logic is straightforward: hand what machines do well to tools, and save human time for the parts that require judgment. By the time something truly needs a human, you already have a reference translation to compare against.
What to Check After Translating
- Is the cue count unchanged: you translated 120 cues, you get 120 back
- Were the first and last timecodes altered: compare the timestamps of the first and last cues
- Line length: common subtitle guidelines recommend no more than 16 characters per line for Chinese and 42 for English; English usually runs longer, so a line that fit before may need to wrap
- Are proper nouns consistent: names, product names, terminology — search the whole file
To make "original + translation" bilingual subtitles, use the bilingual SRT generator: upload an SRT and get a bilingual file directly.
FAQ
Will translation scramble the timeline?
A professional tool will not — the timeline is parsed out as separate data, translation only touches the text, and the file is reassembled on the original timeline. What easily scrambles it is the copy-paste route: once structured data becomes a continuous text block, re-splitting happens with no validation at all.
Is the 500-cue limit on free tools enough?
A 10-minute voiceover video usually has 150 to 250 cues, so 500 cues roughly covers 20 to 30 minutes of content; for anything longer, split by chapter. Note the limit of 3 uses per tool per IP per day.
Can machine-translated English subtitles be published directly?
It depends on the content type. For everyday, information-first speech, translation engines are quite mature. For content involving specialized terminology, brand messaging or legal wording, have a human review it first — the output from the online tool works well as a review draft.