A Conference Recording, Two Caption Files and One Late Fix

Dana runs events for a regional trade association, and in these early autumn weeks she arrived with a ninety-minute recording of a keynote talk. Her captions were already transcribed and timed. What she had not decided was how to hand them over. The talk needed to go on the association's own website, onto a shared video page used by members, and into a short highlight reel for email.

She assumed one file would cover all three. I asked her to open the caption file she had saved and tell me what it was called and what ending it carried. She read out the name, and the ending turned out to be a plain text export that no video player would recognize. That was the real starting point of our session.

I like beginning with the file itself, because the format question is easier to answer once a learner has seen what is actually on the disk.

Both formats are plain text files that pair a block of words with a start and end time. An SRT file numbers each caption, then lists the times, then the text. A VTT file opens with a short header line and uses a slightly different time notation, with a period before the milliseconds where SRT uses a comma.

That small difference matters more than it sounds. Some players on the web expect VTT and quietly ignore SRT, while many editing programs and older platforms take SRT without complaint. VTT also allows a little extra, such as positioning a caption higher on the screen when something important sits at the bottom.

Dana opened both in a basic text editor so she could see there was nothing mysterious inside. Seeing the timing lines in ordinary type took away most of her worry. She could tell which file was which just by reading the first few lines.

• SRT: numbered blocks, comma before milliseconds, widely accepted
• VTT: header line, period before milliseconds, common on web players

We went through her three destinations one at a time. Her own website used a player that read VTT, so that file went there. The shared member page accepted SRT uploads, so we exported a second file for it. For the email highlight reel, nothing in an email can load a separate caption file, so the words had to be part of the picture itself.

That third case is where embedded captions come in. Burning text into the video means every viewer sees it, with no menu to switch on. The cost is that viewers cannot turn it off, change its size or have it read by assistive tools, and fixing a typo means exporting the whole video again.

Dana wrote the rule in her own words: separate files when the platform can use them, embedded text only when nothing else can carry the captions. I thought it was a better summary than the one I had prepared.

After the exports, Dana loaded the VTT file into her website player and played the first few minutes. Everything looked right until the speaker mentioned a product name, and the caption showed a different spelling from the one on the association's slides. She had caught it only because she had actually watched the result.

Because the captions lived in a separate file, correcting the spelling took a minute in a text editor and a fresh upload. Had she relied on the embedded version for everything, she would have rendered the whole video again for one word.

This is the practical argument for keeping an editable caption file even when you also need an embedded copy. Keep the master text file, make the burned-in version last, and label each export with where it is meant to go so nobody uploads the wrong one later.

• Keep one master caption file and export from it
• Name each file for its destination and format
• Watch the first minutes on the real player before publishing

By the end of the session Dana had a half page of notes taped beside her monitor. They were not rules from a textbook, just the questions she now asked before any export. Which platforms will carry this video? Can each one accept a separate caption file, and in which format? Does any destination leave no choice but embedded text?

She also added a line about encoding. When her first VTT export showed odd symbols in place of apostrophes, we traced it to the text being saved in the wrong character setting. Saving again as UTF-8 solved it, and she now checks for that on every export.

None of this is advanced. It is a handful of habits that turn a confusing last step into a calm one, and they hold up whether the video is a keynote, an interview or a short clip for members.