Why the name reading breaks auto-captions
A commencement video has two very different halves. The speeches are ordinary sentences spoken slowly into a good microphone. Auto-captions handle them reasonably well. The name reading is a different problem.
- Rare names. Speech recognition guesses common words. A name it has never seen becomes the nearest dictionary word, or two of them.
- Accents and pronunciation. The reader may pronounce a name the way the family does, which is often not how the software expects it to be spelled.
- Crowd noise. Cheers, air horns and applause land right on top of the next name.
- Speed. Names come in fast succession with no sentence around them, so the software has no context to correct itself.
The result is a caption track where the speeches look fine and the part families care about most is wrong. A graduate's name spelled wrong in public is the error people notice and share. That is why this section needs its own workflow, not a quick skim of the auto-captions.
Use the printed program as your caption glossary
You already have the best source for every spelling: the printed program. DCMP's Captioning Key says proper names, technical terms, and specialized language must be verified against an authoritative source, and the program is exactly that. Get the final file from whoever laid it out, not a photo of a printed copy.
- 1
Get the program as text
Ask for the final document or spreadsheet the program was built from. Copy the graduate list into a plain text file in the order names will be read.
- 2
Check the details that get lost
Keep hyphens, apostrophes, accent marks, spaces inside surnames and capital letters in the middle of names. These are the details auto-captions drop first.
- 3
Note last-minute changes
Ask the registrar or front office for any names added, removed or corrected after printing. The program is the glossary, but the final roster wins.
- 4
Add the other names
List the principal, superintendent, board president, valedictorian, salutatorian, guest speakers and the reader of names, with titles as they will be introduced.
- 5
Replace, don't edit
In the name-reading section, paste the correct names over the auto-caption guesses line by line, then adjust timing so each name appears as it is spoken.
Use the same spelling everywhere. DCMP also says to be consistent in the spelling of words throughout the media. That matters more than usual here, because the same name will show up again in a Reel, a Facebook clip and maybe a yearbook video. Your glossary file is what keeps them all matching.
Speaker labels across a ceremony
A ceremony has a cast: the principal opens, the valedictorian speaks, the board president certifies the class, and someone reads names. Viewers who rely on captions need to know who is talking when the speaker changes.
DCMP's guidance is to wait until a speaker is introduced, in the audio or with on-screen text, before naming them. When the caption position can't show who is talking, it says the speaker's name should be in parentheses on its own line. For example, a caption block for a speaker change might look like this:
(Principal Reyes) Good evening, families, staff and the Class of 2027.
For the name reading, label the reader once when they start, then let the names run. You don't need to label every name with the reader's name again. Add sound tags such as [cheering] or [applause] where they matter, but keep them short so they don't crowd out the next name.
When the diploma walk needs audio description
WCAG 2.1 AA asks for audio description of recorded video when the picture shows something the soundtrack doesn't. Our guide on when audio description is required walks through the test. For a graduation, the answer changes from section to section.
- Usually covered by the soundtrack: a speech where the person at the podium is introduced by name, and the diploma walk when each name is read aloud as the graduate crosses the stage.
- Usually needs description: on-screen text that nobody reads aloud, such as a slide of honors or a lower-third with a name; a silent moment like a moment of remembrance with a photo on screen; a surprise visitor on stage; and highlight reels set only to music, where the picture is the whole story.
- Worth a judgment call: cap decorations, honor cords and stoles. Describe them when they carry meaning the ceremony refers to, such as a cord the speaker mentions.
Not sure about a particular video? Try the check below, or the full audio description check.
What the rules ask of each version
Both the ADA Title II rule and the Section 504 rule adopt WCAG 2.1 AA. Under the dates as extended in 2026, the ADA Title II deadline is April 26, 2027 for public entities with a population of 50,000 or more, and April 26, 2028 for smaller entities and special districts. For Section 504, which covers HHS-funded health centers and others, it is May 11, 2027 for organizations with 15 or more employees and May 10, 2028 for those with fewer. Check yours with the deadline checker.
Here is how that plays out for the versions of one ceremony. This is how we read the rules, not legal advice.
| Version | What applies |
|---|---|
| Full ceremony on YouTube, posted after your deadline | In scope. Recorded video needs accurate captions, and description where the picture shows something the soundtrack doesn't. |
| Ceremony video embedded on a current website page | In scope, whenever it was recorded. |
| Reels and Facebook clips posted after your deadline | The DOJ says a public entity's social media posts made after its compliance date probably do need to comply. |
| Social media posts made before your deadline | Exempt. The DOJ's example is a 2017 social media post by a city's sanitation department. |
| Older ceremony uploads on YouTube | Probably exempt (unclear). Neither DOJ nor HHS said whether YouTube counts as social media. |
| Any version someone asks for | Make it accessible, whatever its date. |
The archived content exception covers content made before the deadline, kept only for reference, not changed since, and placed in an area clearly labeled as an archive. A 2022 ceremony still embedded on the graduation page doesn't qualify. Read the archived video exception guide before you assume a page is out of scope. One more point from the DOJ: you don't have to make the whole social media platform itself meet WCAG 2.1. You're responsible for what you post.
Cutting highlight clips without re-captioning
Once the full ceremony has checked captions, every clip you cut from it already has correct text. The mistake is uploading each clip fresh and letting the platform caption it again, which brings back every misspelled name you just fixed.
- 1
Export the caption file
Download the checked captions from YouTube Studio as an SRT file. Our SRT format guide explains each line.
- 2
Trim to the clip
Copy only the caption blocks inside the clip's start and end times into a new file. Shift the timestamps so the first caption starts near zero.
- 3
Upload with the clip
Facebook accepts an uploaded caption file for videos. For platforms that only offer their own auto-captions, use the trimmed file as your reference when you edit.
- 4
Check the names one more time
Play the clip with captions on and compare each name to your glossary file.
How to add captions to Instagram Reels
Instagram can generate captions for a Reel. Turn them on while you edit the Reel, then open the caption editor and correct each line against your trimmed caption file before you post. Menus move between app versions, so check the current steps in the app.
If you can't edit the automatic captions well enough, burn the text into the video in your editing app instead. Open captions are always visible and can't be turned off, which suits short clips that people watch with the sound off. Our post on closed vs open captions covers the trade-offs. Keep the text large, high-contrast and clear of the app's buttons.
For a music-only highlight Reel, captions alone don't tell a listener what happened. Add a short spoken or text introduction of who is shown, or post a described version and link to it.
A one-week plan for a two-person team
- 1
Before the ceremony
Get the program file and final roster. Build the glossary. List speakers in order.
- 2
Day 1
Upload the full ceremony to YouTube as unlisted. Let auto-captions generate.
- 3
Day 2
Person A corrects the speeches and adds speaker labels. Person B replaces the name-reading captions from the glossary.
- 4
Day 3
Swap sections and proofread each other's work. Note any parts that need description.
- 5
Day 4
Add description where needed, then make the video public and update the website embed.
- 6
Day 5
Cut highlight clips, trim the caption file for each, and post them with checked captions.
- 7
Day 6-7
Answer any family requests for corrections or an accessible version, and fix the glossary so next year starts ahead.
Sources
- U.S. Department of Justice, ADA.gov
- U.S. Department of Justice, ADA.gov
- Described and Captioned Media Program (DCMP) Captioning Key
- Described and Captioned Media Program (DCMP) Captioning Key
This is a reading of the rules, not legal advice. Something out of date? Tell us and we'll fix it.