Repair SRT and VTT Files Without Losing Caption Details

A subtitle file can contain all the right words and still fail to import. A malformed timestamp, missing block separator, or wrong export format can stop an application from recognizing it. Repair starts with identifying that structural problem, not deleting everything except dialogue.

Keep the original file, including any editor project that produced it. Open a copy in the Motionix subtitle fixer, review its diagnostics, and export only after you understand what will change. A readable output is not necessarily a faithful replacement for the source.

Recognize the format you actually have

A typical SRT block contains a sequence number, a timing line using comma-separated milliseconds, then text. Blank lines separate blocks:

1
00:00:05,200 --> 00:00:07,800
The meeting starts here.

WebVTT starts with a WEBVTT header followed by a blank line. Its timestamps use a period before milliseconds; cue identifiers are optional:

WEBVTT

00:00:05.200 --> 00:00:07.800
The meeting starts here.

Changing .srt to .vtt in a filename does not convert the contents. Conversely, a file named .txt may contain valid subtitle text, but the destination application may still require a supported extension. Choose the actual output format and confirm the downloaded file's contents.

Separate errors from review warnings

Check timestamps for valid minutes and seconds, an end later than the start, and the correct timing arrow. Check that an accidental blank line has not separated a timing line from its text. For WebVTT, use UTF-8 and keep cue starts in nondecreasing order.

An overlap is different. WebVTT subtitle cues may overlap intentionally; chapter tracks have different constraints. Simultaneous speakers or a continuing sound description can require text at the same time. Do not shorten every cue just to make a warning disappear.

Likewise, a long line is a readability concern rather than automatic proof of invalid syntax. Review wrapping in the destination player. Preserve meaningful line breaks, but do not insert empty lines inside a single cue because they terminate the block.

Read the loss report before conversion

WebVTT can carry NOTE, STYLE, and REGION blocks, cue identifiers, positioning settings, voice or language tags, and timestamps inside cue text. Some files carry timed metadata rather than subtitles. These features are not all interchangeable with basic SRT, and a simple repair workflow may not support them.

Read every unsupported-feature or export-loss warning. If a feature is rejected, return to the original editor or produce an explicitly simplified copy there. If an export omits information, keep the original and decide whether that loss is acceptable for this particular delivery. Never assume conversion preserves everything because the spoken words remain.

Pay special attention to speaker identification and accessibility information. Removing a voice tag without retaining a visible speaker label can change meaning. Removing inline timestamps may destroy karaoke-style timing even when the outer cue times survive.

Verify the delivered file

Compare cue counts, the first and last cues, accented characters, multiline captions, and a passage with simultaneous speech. Investigate any omitted blocks rather than accepting a lower count as cleanup. Reopen the export in the actual destination application and watch representative sections with sound.

If the file now imports but text is early or late, follow offset and drift correction. Structural repair does not automatically synchronize captions, correct transcription, or repair the video itself.

Sources and review limits

Reviewed November 6, 2026. MDN's WebVTT guide documents header, encoding, cues, and permitted overlaps. The W3C WebVTT Candidate Recommendation Draft covers richer syntax and parsing; it is a work in progress, not a certification of Motionix. These references do not establish uniform SRT support across editors, and no all-editor compatibility test is claimed.