Choose SRT when your destination asks for an SRT caption file, and WebVTT when your website or player requires VTT text tracks. Both store timed text separately from the video. The right choice starts with the receiver's requirements, not a claim that one format makes captions more accurate.
For a typical handoff, preserve an approved caption master, make the requested delivery copy, and test that copy with the final video. Changing a filename from .srt to .vtt is not a complete conversion. The file contents and the receiving player's behavior matter.
Oakgen Compose provides an SRT download for captions on its subtitle timeline. For a player that needs WebVTT, export the SRT from Compose and follow the SRT-to-VTT conversion guide. The guide also covers shifting cues when the captions are consistently early or late.
SRT and VTT at a glance
The Library of Congress format description describes SRT as a text subtitle sidecar, rather than a video or audio file. The W3C WebVTT specification defines timed text resources for web media, including captions and subtitles.
| Question | SRT | WebVTT |
|---|---|---|
| Typical filename ending | .srt | .vtt |
| Familiar cue layout | Number, timing line, text, blank separator | Header, then timing/text cue blocks; identifiers are optional |
| Millisecond separator in timestamps | Comma, such as 00:00:01,200 | Period, such as 00:00:01.200 |
| Required file signature | No WEBVTT header | Begins with WEBVTT |
| Styling expectation | Keep delivery simple; support varies by application | Format supports cue settings and styling, subject to receiver support |
| Best selection rule | Deliver it when requested and test ingestion | Deliver it when requested and test playback |
Neither extension proves that the text is correct, synchronized or accessible. An accurately formatted file can still contain the wrong name, omit a meaningful sound or appear several seconds late.
See the same cue in both formats
Here is a small invented example for a product walkthrough. It is not a transcript from a tested video.
SRT:
1
00:00:01,200 --> 00:00:03,400
Choose the smaller tray.
2
00:00:03,700 --> 00:00:05,800
It fits inside the drawer.
WebVTT:
WEBVTT
00:00:01.200 --> 00:00:03.400
Choose the smaller tray.
00:00:03.700 --> 00:00:05.800
It fits inside the drawer.
Notice the header and punctuation changes. The words and display intervals are the same. For this simple case, conversion need not rewrite the message or shift its timing. Files with styling, speaker markup or other features deserve a closer inspection than this two-cue example.
MDN's WebVTT guide documents UTF-8 encoding, the header, cue blocks and optional settings. If a developer is connecting your captions to a website, give them the intended language and caption kind as well as the VTT file. A subtitle file sitting beside an MP4 is not automatically connected to a web player.
First decide: sidecar captions or text burned into the video?
A sidecar is a separate timed-text file. Burned-in captions become part of the rendered picture. Changing SRT to VTT addresses the sidecar format; it does not burn text into an MP4.
Use a delivery brief like this before exporting anything:
Final video: drawer-demo-approved-04.mp4
Language: English
Caption purpose: speech plus meaningful non-speech information
Required delivery: separate subtitle file, receiver requests VTT
Additional delivery: social video with visible captions
Approval source: corrected caption master for video revision 04
Playback check: beginning, middle, ending, plus every edited passage
Those are two different deliverables. If you send only the social video, the website team may still need a separate caption file. If you send only the sidecar, a social posting workflow may not display it automatically. Confirm the actual upload path and requested format rather than assuming every platform accepts the same package.
The W3C captions resource distinguishes captions that viewers can switch on and off from open captions that remain visible. It also explains why relevant non-speech information can matter. Include what a viewer needs to understand the content, rather than treating automatic dialogue transcription as finished caption authoring.
A practical conversion and handoff workflow
1. Freeze the video revision
Choose the exact video the receiver will use. If someone later removes an opening pause, inserts an intro or changes playback speed, captions aligned to the older file may no longer match.
Record the final filename and duration alongside the caption master. Version labels do not fix timing, but they make it easier to discover that a receiver loaded the wrong pair of files.
2. Correct the caption master
Read for names, technical terms, punctuation, speaker changes and important sounds. Then watch with the captions enabled. Reading the file alone cannot reveal whether a cue arrives before or after the relevant sound.
An editorial check is still necessary when automatic transcription is used. Keep correction decisions in the master rather than editing three delivery copies independently. Otherwise the next conversion can reintroduce an old spelling mistake.
3. Convert a copy with the receiver's requirements in view
Use a converter or editor whose supported input and output formats you have checked. For a simple SRT-to-VTT conversion, inspect the resulting header, timestamp punctuation, cue boundaries and encoding. For VTT-to-SRT, identify any positioning or styling that may not survive the receiving application.
Do not perform a global comma replacement across the entire document: dialogue can contain commas too. A format-aware conversion should handle timing syntax without rewriting ordinary punctuation. Likewise, do not remove lines merely because they look unfamiliar; first determine whether they are cue identifiers, settings or content.
4. Compare the text and timing
For a straightforward sidecar conversion with no intentional restructuring, compare the first cue, the last cue and a sample from the middle. Check that the wording remains intact and that display intervals match the master. Investigate unexpected changes in cue count rather than assuming they are harmless.
Search for a distinctive phrase near the end to catch accidental truncation. Test accented characters or other scripts used in the actual captions. A file that looks correct in one text editor may still fail the receiver's ingestion or rendering expectations.
5. Test the delivered pair
Load the final video and delivery subtitle file into the intended receiver or its review environment. Verify the selected language, start timing, readability and the end of the video. Test both the presence of captions and the option to hide them when a switchable track is required.
If the receiver cannot provide a preview, mark playback verification pending in the handoff. Do not describe a conversion as fully verified because it saved without an error.
Where Compose fits
In the current Oakgen workflow, Compose has a Subtitles panel and can download the existing subtitle timeline as an SRT file. Automatic subtitle creation is a credit-based operation for eligible paid accounts. Editing the video and generating new transcription are separate actions; review the quote before starting transcription.
Compose export access also depends on the purchased offer. Current offer rules include Ultimate/Creator access with legacy exceptions, so opening an editor does not establish export entitlement for every account. Check pricing and the account's available export action. No free or unlimited generation is implied.
A sensible handoff is therefore:
- Work with the correct video revision in Compose.
- Review and correct the subtitle text and timing.
- Download the SRT for that approved timeline.
- If the receiver requires VTT, convert a copy outside this documented SRT-download path.
- Check the delivered subtitles against the final video in the receiving player.
This workflow does not establish a VTT exporter or sidecar importer in Oakgen. If your task depends on those functions, verify the current interface before planning around them.
Troubleshoot the right layer
The file is rejected immediately. Check the requested extension, actual text syntax, encoding and receiver requirements. A renamed file may still contain the original format.
Captions load but drift later. Confirm that the video and captions share the same revision. Check whether the video was retimed or cut after the caption master was approved. Changing formats alone is not a timing repair.
Styling disappears. Separate caption content from presentation. A receiver may ignore styling or offer its own controls. Ask for its supported features and validate there instead of assuming the source editor's appearance travels with the file.
There are two layers of captions. Check whether the MP4 already has visible text and the player is also displaying a sidecar track. Avoid duplicating the same message unintentionally.
For related preparation work, see the voiceover timing guide and audio sample-rate handoff. They solve timing and audio preparation problems, rather than choosing a subtitle container.
Ready to organize the edit and its subtitle master? Create an Oakgen account, inspect Compose, and confirm your plan and credit conditions before running paid actions. Deliver the file the receiver needs, with the approval and playback limits clearly recorded.


