That's the short answer. The longer one covers what goes into a good caption track, how it travels with a TV broadcast or a web video, who is legally required to provide it, and how to make one for your own video in a few minutes.
What closed captioning includes
A transcript tells you what was said. Closed captions tell you what a hearing viewer would have heard. So a proper caption track has three kinds of content.
First, the dialogue, word for word, in the order it was spoken. Second, who is speaking, whenever that isn't obvious from the picture: a narrator, a voice on the phone, someone off-screen. Third, the sounds that carry meaning. A phone ringing, a door slamming, a laugh track, a song with lyrics, a tense score swelling before the jump scare. If a hearing viewer would react to it, the caption should mention it.
The W3C's accessibility guidance puts it the same way: captions convey spoken dialogue plus "sound effects, music, laughter, speaker identification and location." That third category is the part people forget, and it's the main thing that separates captions from plain subtitles. There's more on that in our guide to how captions differ from subtitles and SDH.
Why is it called closed captioning?
Because the captions are closed, meaning hidden, until someone asks to see them. The text travels with the video as separate data, and the TV or player only draws it when the viewer switches it on.
In the early days this was literal. The caption data was tucked into part of the TV signal you never see, and you needed a decoder box to reveal it. Without the box, the picture looked normal. Today the decoder is built into every TV, phone and browser, but the name stuck.
The opposite is open captions: text burned into the picture, visible to everyone, impossible to turn off. Each has its place, and we compare them in detail in our guide to open and closed captions.
How closed captions are delivered
On broadcast TV
Analog TV in North America carried captions on Line 21 of the vertical blanking interval, a strip of the signal just above the visible picture. That format is known as CEA-608. Digital TV uses CTA-708 (formerly CEA-708), which carries the old 608 captions for backward compatibility plus up to 63 additional caption streams. The caption data rides inside the broadcast itself, so it reaches viewers along with the picture and sound.
708 also gave viewers control over the look. FCC rules require digital TV caption decoders to let you scale the text from 50% to 200% of the default size and pick the font, text color and background color and opacity.
On streaming services and the web
Online video usually keeps captions in a separate text file, often called a sidecar file, that sits next to the video. The two you'll
meet most are SRT and WebVTT (.vtt). Both are
plain text: a start time, an end time and the words to show. On a web page, a VTT file is attached to an HTML video with a
<track kind="captions"> element, and the browser adds a CC control.
YouTube accepts a long list of formats, including SRT, VTT and the broadcast format SCC, which carries CEA-608 data. That's handy if you're moving a TV program online. For everyone else, an SRT file is the simplest thing to make and upload.
The closed caption symbol
The closed captioning logo most people know is the white "CC" inside a rounded TV-screen shape. It was designed in the early 1980s by Jack Foley, a senior graphic designer at the Boston public broadcaster WGBH. At the time, the only caption symbol around belonged to the National Captioning Institute (NCI), whose "TV with a tail" logo, a TV set merged with a speech balloon, could only be used on programs NCI captioned. WGBH made its own for the shows it captioned, then released it into the public domain in the mid-1990s so anyone could use it.
You'll see the closed caption symbol in TV listings, on DVD cases and on the caption button in most video players. On YouTube and many other players, the button just says CC.
Closed captioning examples
Here's a short scene captioned the way professional caption writers would do it, written as an SRT file you could upload today:
1
00:00:01,200 --> 00:00:03,400
MAYA: Did you hear that?
2
00:00:03,900 --> 00:00:05,300
[floorboard creaks upstairs]
3
00:00:05,800 --> 00:00:08,600
DAD (off-screen): It's just the house settling.
Go back to sleep.
4
00:00:09,400 --> 00:00:12,000
[slow, tense music]
5
00:00:14,200 --> 00:00:17,500
♪ Happy birthday to you ♪
6
00:00:17,600 --> 00:00:19,000
[everyone laughs]What these closed captions examples get right: the speaker is named when the viewer can't see who's talking (cue 3). Sounds go in square brackets and describe what matters, not every rustle (cues 2 and 6). Music without words is described by mood (cue 4), and sung lyrics sit between music notes (cue 5). Each cue is short enough to read before it disappears.
What bad captions look like is just as instructive: one giant block of text per sentence, no speakers, and a shriek captioned as nothing at all. If you only fix one thing in an auto-generated file, add the sound cues.
A short history of closed captioning
Closed captions were first demonstrated in the US in December 1971, at a national conference on television for hearing-impaired viewers in Knoxville, Tennessee. In February 1972, ABC and the National Bureau of Standards showed captions hidden in a broadcast of The Mod Squad at Gallaudet College. That same year, PBS began open-captioned broadcasts of The French Chef. In 1976 the FCC set aside Line 21 for caption data.
Regular closed captioning started on March 16, 1980, with The Wonderful World of Disney, The ABC Sunday Night Movie and Masterpiece Theatre, captioned by NCI. To see the captions you had to buy a decoder box, which Sears sold for $250.
That price was the real barrier, and Congress fixed it with the Television Decoder Circuitry Act of 1990. From July 1, 1993, TV sets with screens 13 inches or larger sold in the US had to be able to display closed captions on their own. Digital TVs of that size must now include CTA-708 decoders.
Who requires closed captioning
In the US, the FCC's closed captioning rules apply to TV programming. Since a 2012 rule under the 21st Century Communications and Video Accessibility Act (CVAA), programs that aired on TV with captions must also be captioned when they're shown online, and in 2014 the FCC extended this to clips of those programs. Video that has only ever been online isn't covered by these rules.
For websites, the reference point is the Web Content Accessibility Guidelines. WCAG Success Criterion 1.2.2 requires captions for all prerecorded audio in video, at Level A, the most basic level. W3C accepts open or closed captions for it.
The Americans with Disabilities Act (ADA) now points to WCAG too, at least for government. A 2024 Department of Justice rule requires US state and local governments to make their web content and apps meet WCAG 2.1 Level AA, which includes 1.2.2. After an extension in April 2026, the deadlines are April 26, 2027 for larger governments and April 26, 2028 for smaller ones. Private businesses aren't covered by that rule, but many universities and companies set the same bar in their own policies.
What makes closed captions good
In 2014 the FCC adopted quality standards for TV captions. They're written for broadcasters, but they work as a checklist for anyone:
- Accurate: the words match what was said, and background sounds are conveyed as fully as possible.
- Synchronous: each caption appears when the words are spoken and stays up long enough to read.
- Complete: captions run from the start of the program to the end, not just the first few minutes.
- Properly placed: captions don't cover faces, on-screen text or other important parts of the picture.
Placement is the one creators miss most. If your video has a name title or a chart at the bottom of the frame, a player that draws captions in the default spot will cover it. Leave that area clear, or move the caption up for those lines if your format allows it.
How to add closed captions to your own video
You need a caption file, and then a place to upload it. The fastest route is our SRT generator: drop in your video, and it writes a timed transcript you can correct line by line. The video is opened in your browser and isn't uploaded; only a compressed audio track goes out for transcription. Then download an SRT or VTT.
One honest limitation: our auto captions tool writes what's said. It doesn't label speakers or describe sounds, so type "MAYA:" or "[door slams]" into the lines where they matter before you export. That small edit is what turns a transcript into closed captions.
To put the file on YouTube, open YouTube Studio, choose Subtitles, pick the video, add a language, then click Add and Upload file (with timing). On your own site, attach the VTT with a track element. Already have a caption file with typos or drifting timing? Open it in the subtitle editor, with or without the video, to fix the text, adjust individual times or shift every line at once.
And if you'd rather rely on YouTube's own captions, read how YouTube auto-captions work first. They're a decent start, but they won't add the sound cues for you either.