Caption files are fragile in the best pipelines. Producers lock picture, localization receives SRT masters, then editorial cuts another two seconds from the head for the social cut. Every stakeholder still expects accessible subtitles on the final MP4. Teams that treat time shift as a standard step ship corrected SRT faster than reopening desktop subtitle suites for hundreds of cues.
Editors: after the trim lands
Picture lock is rarely the last export. Common late changes include:
- Removing a sponsor read or disclaimer
- Swapping a stock intro for a brand sting
- Tightening dead air before the first question
- Splitting one long master into a YouTube version and a course module
When dialogue text is unchanged, the fix is arithmetic, not retranscription. Measure how far the timeline moved, then run the master SRT through the SRT Time Shifter with that offset.
Premiere and DaVinci Resolve handoffs
- Export captions as SRT from the previous timeline version (keep the file named with the version number).
- Note the duration removed at the head or inserted before the first spoken line.
- Shift the SRT in the browser tool.
- Import the shifted SRT on the new timeline and spot-check five cues across the runtime.
- Archive both the unshifted and shifted files in the project folder.
If you maintain separate sequences for :30 social cuts, duplicate the SRT and apply a different offset per sequence rather than editing cue times manually in three places.
When drift shows up late in the program
Long interviews and webinars sometimes exhibit slow sync slip instead of a single jump. Editors hear it before viewers complain: the guest is clearly mid-sentence but the caption track lags.
Try a time scale slightly above or below 1 in the SRT Time Shifter, then combine with a small offset so the opening still lines up. Document the scale value in the project readme so the localization vendor can match it on translated tracks.
Agencies: client review without caption rebuilds
Agencies often deliver:
- A review MP4 with burned-in captions for stakeholders who will not open an NLE
- A clean SRT for the client’s internal LMS or YouTube channel
When the client asks for “just take five seconds off the top,” the agency should not re-bill a full re-caption pass if the words are identical. Shift the approved SRT, re-burn if needed, and note the offset in the delivery email so the client’s team can reproduce the same fix on their side.
For multi-episode retainers, store a caption changelog next to each SRT:
| Version | Offset (s) | Scale | Notes |
|---|---|---|---|
| ep03-v2 | +4.5 | 1 | Added logo sting |
| ep03-v3 | +4.5 | 1.005 | Drift fix after 22:00 |
That table prevents translators from working against an old master.
Localization: keep languages aligned
Localization teams usually receive one timed reference (often English) and several translated SRT files. If editorial shifts the reference by a constant offset, every language file needs the same shift unless dubbing changed pacing.
Batch workflow:
- Confirm dialogue text is still valid in each language.
- Upload all
.srtfiles to the SRT Time Shifter with one offset and scale. - Download the ZIP of shifted files.
- QA one cue per language at the head, middle, and tail of the video.
If a dubbed track no longer matches the original timing, text may need edits. Use SRT to TXT for a quick read-through, adjust copy, then TXT to SRT only for the sections that changed. Reserve time shifting for when words still match the audio.
Accessibility and compliance teams
Accessibility reviewers care that captions track speech, sound effects where required, and speaker changes. A global offset does not remove the need for accurate transcription, but it does fix systematic early/late errors that fail spot checks after an editorial change.
Document shifted files in your ticket system:
- Original SRT hash or filename
- Offset and scale applied
- Player used for verification (web, LMS, embedded iframe)
That audit trail helps when a regulator or enterprise customer asks how you validated captions after a late trim.
Course and corporate video libraries
Training libraries reuse modules for years. A module re-exported at a lower bitrate should not force authors to retime captions from scratch if the audio waveform is unchanged. Shift once, upload the new SRT to the LMS caption field, and keep the offset in the module metadata for the next refresh.
When narration script changes, rebuild captions from text rather than shifting alone. The time shifter assumes cue text still matches spoken audio.
Browser-based shifting for sensitive drafts
Draft compliance training, unreleased product demos, and executive town halls often cannot ride through unknown upload services. Shifting in the browser keeps SRT files local while editors and vendors iterate on sync.
Limitations to plan around
- Frame rounding: Players snap timestamps to frame boundaries. Expect to nudge by a fraction of a second after import.
- Non-linear edits: Removing a section from the middle of a video requires splitting the SRT or re-exporting from the NLE, not a single global offset.
- Burned-in captions: Shifting SRT does not change pixels already baked into video. Re-burn or deliver a sidecar file.
Related: Walk through offset and scale step by step in How to shift SRT subtitle timing online.