SRT Subtitle Format: Fix Timing, Encoding, and Import Errors

Read one cue as four separate parts

The Library of Congress format description documents this basic structure and notes that SRT is not rigorously standardized. Different applications can vary in what they tolerate, especially around formatting. Use the simplest supported structure when moving between tools.

A cue may contain more than one line of text. A line break inside a cue changes its visual grouping; an empty line separates cues. Do not insert a blank line merely to create a bigger gap between two lines that belong to the same cue. That can change how the parser interprets the file.

The file contains timing and text, not the video itself. Naming it similarly to the video can help organize a handoff, but the destination may still require you to import or select it explicitly. Keep the media and its caption file together with a version note.

Open the valid file before repairing the broken one

Download the valid SRT Download the intentionally broken SRT Download the repair checklist

Both files use an original short teaching transcript about preserving a recording. The valid version contains three cues. Its timings are illustrative, not aligned to a spoken recording. Use it to inspect file structure and import behavior; replace the timings when you have real media to caption.

1
00:00:01,200 --> 00:00:03,600
Keep the original recording.

2
00:00:04,000 --> 00:00:06,500
Save a separate edited copy.

3
00:00:07,000 --> 00:00:09,500
Check the final export.

The broken version changes the first arrow to a single hyphen, reverses the second cue's start and end, and duplicates a cue number. Each mistake has a different repair. A permissive application might accept part of that file, so “it imported” is not enough evidence that all three cues survived correctly.

Work on a copy and correct one issue at a time. Compare the result with the valid reference, then reopen it in the application you intend to use. Our local check verifies the expected syntax and detects the planted defects. It does not claim acceptance by every caption editor or streaming service.

Repair syntax before changing timing decisions

Use a plain-text editor or a subtitle editor. A word processor can add document formatting even when the page looks like ordinary text. Also check that the filename really ends in .srt, rather than .srt.txt with the last extension hidden by the operating system.

Do not apply an unrestricted search-and-replace to every comma or period in the document. Timestamp punctuation and sentence punctuation serve different jobs. Change the timing fields deliberately, then inspect the spoken text for accidental edits. Keep a copy of the original failed file until the repair is confirmed.

Use the destination's encoding and styling rules

YouTube's supported caption formats documentation specifies plain UTF-8 for basic SRT files and says styling markup is not recognized for that format. A colored or positioned caption in one editor can therefore lose that appearance when imported through another service.

Check words containing accents, apostrophes, and non-Latin characters after saving and reopening the file. A successful save message cannot tell you whether the intended characters survived. If the application asks for an encoding, record the selected value in the delivery note.

Separate content from appearance. If the handoff needs precise visual styling, establish whether the destination supports a suitable caption format or whether a deliberately burned-in version is required. Keep a clean editable caption file as well when the workflow allows it. Burning captions into pixels makes later corrections more costly.

A WebVTT file is not an SRT file with a different extension. Its syntax includes differences such as the header and timestamp conventions. Use a supported conversion workflow and inspect the result instead of renaming the file and hoping the player guesses correctly.

Distinguish a fixed offset from a changing mismatch

If every cue appears half a second late, a consistent timing shift may solve the problem. If the first cue is correct but later cues drift progressively, shifting everything by one constant amount will not fix the whole file. Check whether the caption track belongs to a different edit or a different-duration version of the media.

Choose an early, middle, and late spoken phrase and compare the actual onset with its cue. Write the differences down. If the error starts suddenly after a cut, inspect that edit first. A removed introduction, inserted pause, or alternate take can invalidate only part of a previously correct caption track.

For example, removing a two-second title card from the beginning changes every later time relative to the new video. Removing a pause in the middle changes only the later cues. Preserve the original subtitle file and perform the correction on a version explicitly matched to the new cut.

The sync diagnosis guide explains offset and drift in more detail. If your transcript is correct but timings have not been created, the forced-alignment workflow covers a starting point for assigning word times. Neither process eliminates final caption review.

Check readable phrases, not just valid timestamps

A technically valid cue can still be difficult to read. Review phrase breaks, display duration, and how much text appears at once. Keep a person's name together where practical, and avoid line breaks that briefly change the meaning. Judge the rendered caption at ordinary viewing size rather than only reading the source file.

Do not treat a fixed character count as a universal law of SRT syntax. Line length and reading-rate targets belong to the destination's editorial guidance and the intended audience. They are useful review criteria, but they are different from whether the timestamp can be parsed.

If the content includes meaningful non-speech audio, a dialogue-only transcript may leave viewers without necessary context. The SDH caption guide shows how speaker identification and sound cues change an otherwise similar caption track. Keep that content decision separate from the file-format repair.

Finish with an export-and-playback check

1. Match the subtitle filename and version to the final media.
2. Reopen the saved SRT and inspect its first and last cues.
3. Import it into the actual destination or intended player.
4. Check early, middle, and late speech, including a cue after any major edit.
5. Confirm the caption language, label, and visibility controls.
6. Retain the approved file and a note of what was checked.

If the video was re-exported after the captions were approved, compare duration and edit structure again. A color correction may leave timing unchanged; a trimmed opening will not. For a video podcast, treat the full episode and its short clips as separate caption deliveries.

Frequently asked questions

Start with a checked narration

Prepare an approved narration using available Voice Lab options, then create and review its timed captions in a caption editor.

Open Voice Lab