Subtitle Gap Analyzer

The Subtitle Gap Analyzer measures the silent interval between the end of one cue and the start of the next across the SRT you paste into the SRT Content box. Any interval longer than Max gap before warning (ms) is reported with the two cue numbers and the gap length in seconds; ticking Auto-close large gaps also returns an adjusted SRT with those gaps pulled in.

Big holes in a caption track feel like something has broken. Viewers who rely on captions cannot tell a genuine pause from a missing translation, so long empty stretches read as dropped content. It is common in ASR-derived files where music was skipped, in translated tracks where a section was never delivered, and in edits where a scene was cut but the surrounding cues were not retimed. E-learning producers, documentary subtitlers and QC reviewers sweep for these holes before a client finds them.

The tool compares each cue's start with the previous cue's end and flags the difference when it exceeds your threshold. Auto-close merges nothing: it moves the later cue's start to the previous end plus exactly the threshold, so a 16 second hole at a 2000 ms setting becomes a 2.00 second gap while the cue keeps its original end. If that start would meet or pass the cue's own end, the end is moved out by a millisecond. Because pulling a cue earlier changes its sync, auto-close suits gap-tolerant material rather than tightly dialogue-locked tracks.

Analysis runs entirely in the browser with no upload, and adjusted files download as .srt.

• Configurable warning threshold: The Max gap before warning (ms) field sets what counts as too long, so you can sweep loosely at 5000 ms for missing sections or tightly at 1000 ms for pacing work, and it defaults to 2000 ms when left empty.
• Precise gap report: Each finding names the cue on either side of the hole and gives its length in seconds to two decimal places, making it easy to jump to the right timecode in your editor.
• Optional tightening pass: Ticking Auto-close large gaps appends an adjusted SRT under the report, while leaving it clear keeps the run purely diagnostic.
• Closes to your threshold, not to zero: Auto-close rewrites each oversized gap to exactly the millisecond value you set, so captions stay separated by a deliberate, consistent interval instead of running together.
• End times preserved: Only the start of the later cue is moved, so the cue keeps the out point it was given and its own duration is never silently shortened below validity.
• Degenerate cue protection: If an adjustment would leave a cue ending at or before it starts, the end is pushed out so the block stays legal.
• Cues are never merged or removed: The cue count and all subtitle text come through unchanged, which keeps the file comparable against the source script.
• Instant and local: Even long feature files are analysed in the browser without uploading anything.

1. Paste your subtitle file into the SRT Content box, or press Load sample to work with a file that already contains an obviously oversized gap.
2. Enter a threshold in Max gap before warning (ms), using something like 2000 for ordinary pacing checks or 8000 when you are hunting for whole missing sections.
3. Leave Auto-close large gaps unticked for the first run so you can read the diagnosis before anything is changed.
4. Press Run and review the report, which lists each oversized gap with the cue numbers on either side and its length in seconds, or confirms that no large gaps exist.
5. Check the flagged spots against the video, because a genuine silence should be left alone while a hole caused by missing translation needs new cues rather than tightening.
6. If the gaps really are just slack, tick Auto-close large gaps, run again, and inspect the adjusted SRT that appears below the report.
7. Copy the result or download it as a .srt file, then replay a flagged moment to confirm the retimed cue still lands with the dialogue.

Example 1 - a hole where a section was never translated. With Max gap before warning (ms) at 2000 the analyzer reports the empty stretch so you can go and write the missing cues.

1
00:00:01,000 --> 00:00:04,000
The survey started in March.

2
00:00:20,000 --> 00:00:23,000
By June we had our first results.

Report: Between cue 1 and 2: 16.00s gap.

Example 2 - tightening slack with auto-close on. The same file with the checkbox ticked pulls cue 2 in to sit exactly 2000 ms after cue 1 ends, keeping its original out point.

2
00:00:06,000 --> 00:00:23,000
By June we had our first results.

The second example shows why auto-close is a pacing tool rather than a repair: the caption now runs much longer on screen, so use it where cues are not locked to speech.

• Finds missing content fast: An unexplained dead stretch in a caption track usually means untranslated or undelivered dialogue, and a threshold sweep surfaces it in seconds instead of a full playback review.
• Keeps viewers oriented: Caption users cannot hear whether silence is intentional, so removing needless empty space keeps the reading experience feeling continuous and deliberate.
• Diagnose before you commit: Running report-only first means you never retime a cue that was correctly parked over a genuine pause.
• Consistent rhythm across a series: Using the same millisecond threshold on every episode gives a house pacing standard that reviewers can actually check against.
• Fewer QC notes coming back: Catching oversized gaps in-house avoids the round trip of a client flagging them after delivery.
• Confidential by design: The whole analysis runs in your browser, so scripts under embargo are never uploaded anywhere.