A vendor delivers your captions. The file is clean, accurate, and perfectly timed.
Your broadcast playout rejects it.
Not because the captioning was wrong. Because you needed 708 captions in an MCC file and received an SRT built for web. Same words. Wrong type. Two days lost.
That is the part nobody explains when they say “add captions.” Captions are not one thing. They come in styles, standards, and formats, and picking the wrong combination breaks delivery long after the writing is done.
Why “Just Add Captions” Is Not a Spec
Ask three people what closed captions are and you get three answers. One describes how they look on screen. One describes the broadcast standard. One names a file extension.
All three are correct. They are describing different layers of the same thing.
That confusion is expensive. It produces rework, failed platform QC, and compliance gaps that only surface during an audit. Getting the vocabulary right up front is the cheapest quality control available.
The Three Axes That Define Caption Type
Every captioning decision sits on three independent axes.
- Display style. How text appears and moves on screen. Pop-on, roll-up, or paint-on.
- Technical standard. How caption data is encoded and carried. CEA-608 or CEA-708.
- Delivery format. The file your destination platform accepts. SCC, MCC, SRT, WebVTT, TTML, and others.
A single project routinely mixes all three. Pop-on style, 708 standard, MCC file for broadcast, plus a WebVTT version of the same captions for the streaming cut.
Display Styles: Pop-On, Roll-Up, and Paint-On
The Described and Captioned Media Program defines three display styles in its Captioning Key guidance.
| Style | How it behaves | Best for | Trade-off |
| Pop-on | Blocks appear, hold, then clear as the next block appears | Prerecorded film, drama, streaming, web video | Requires prep time, not viable live |
| Roll-up | Lines scroll upward continuously, typically two to four lines at a time | Live news, sports, real-time events | Trails the speaker slightly, less placement control |
| Paint-on | Characters appear left to right as spoken | Late additions, on-screen text callouts, some live use | Can feel busy for long dialogue passages |
Pop-on is the default for anything prerecorded, because you control exactly where each block sits and can move it away from on-screen graphics. Roll-up exists because live captioning has no prep window. Paint-on is the specialist option, used sparingly.
If you take one thing from this section: display style is driven by whether the content is live, not by preference.
Technical Standards: CEA-608 vs CEA-708
These are the two broadcast caption standards, and they are frequently confused.
CEA-608, also called Line 21 or EIA-608, was built for analog television. Monospaced white text on a black box, four caption channels, almost no styling control. It persists because legacy infrastructure still carries it.
CEA-708, also called CTA-708 or DTVCC, is the digital television standard. It supports multiple fonts, sizes, colours, background opacity, and precise positioning. It also carries 608 data inside it for backward compatibility, which is why teams often deliver both from one workflow.
For broadcast delivery in the US, 708 is the working standard. For streaming, neither applies directly, because OTT platforms use their own text formats instead.
Closed, Open, SDH, and Subtitles
Different category entirely. This axis is about who sees the text and what it contains.
| Type | Viewer can toggle | Includes non-speech audio | Typical use |
| Closed captions | Yes | Yes, sound effects and speaker ID | TV broadcast, OTT, compliance delivery |
| Open captions | No, burned into the picture | Yes | Social feeds that autoplay muted, public displays |
| SDH subtitles | Yes | Yes | Streaming and disc, where 608 and 708 are unsupported |
| Standard subtitles | Yes | No, dialogue only | Translated versions for hearing audiences |
The distinction that trips teams up most: subtitles assume the viewer can hear and only need the language. Captions assume the viewer cannot hear. That is why a translated subtitle file does not satisfy an accessibility requirement.
SDH exists to bridge them, delivering caption-level information through a subtitle-style delivery path. Our guide to translation and subtitle services covers where SDH fits in a multilingual plan.
Caption File Formats and Where They Belong
| Format | Carries | Common destination |
| SCC | 608 data | Broadcast, legacy playout, some archive |
| MCC | 708 data | Digital broadcast delivery |
| SRT | Plain timed text, minimal styling | Web video, social platforms, YouTube |
| WebVTT | Timed text with basic styling | HTML5 players, HLS streaming |
| TTML / IMSC | Rich styling and positioning | Major OTT platforms, DASH delivery |
| EBU-STL | Timed text | European broadcast |
One caption pass should feed all of these. If your workflow recreates work per format, that is a process problem rather than a captioning problem.
Which Type Fits Which Workflow
Work backwards from distribution, never forwards from production.
- Live news or sports on broadcast: roll-up style, 708 standard, MCC delivery
- Prerecorded series for OTT: pop-on style, TTML or IMSC delivery, SDH for the accessibility track
- Social media teasers: open captions burned in, since most feeds autoplay muted
- Corporate or e-learning video: pop-on style, WebVTT or SRT, depending on the player
- Archive digitization: whatever the original carries, normalized to a modern master format
Map every destination before commissioning the work. The captions are the same. The packaging is not.
What the FCC Actually Requires
Style and format are your choice. Quality is not.
The FCC holds captions to four quality standards: accuracy, synchronicity, completeness, and placement. Accuracy covers proper nouns, numbers, and slang. Synchronicity means captions track the audio. Completeness means the full programme, start to finish. Placement means captions do not obscure other essential on-screen information.
Live programming is held to a best-efforts standard rather than the same absolute bar, which is precisely why roll-up captioning is accepted there. Our FCC caption rules explainer breaks down how those obligations apply across TV and streaming.
Mistakes Worth Avoiding
- [ ] Delivering SRT where the platform requires TTML or IMSC
- [ ] Using pop-on captions for live content, which cannot work
- [ ] Treating translated subtitles as an accessibility deliverable
- [ ] Ignoring placement, so captions cover lower thirds and scoreboards
- [ ] Captioning to one destination, then rebuilding for the next
- [ ] Skipping speaker identification in multi-speaker scenes
- [ ] Assuming 608 delivery is still sufficient for digital broadcast
Closed Caption FAQs
What are the three types of closed captions? By display style, they are pop-on, roll-up, and paint-on. Pop-on suits prerecorded content, roll-up suits live programming, and paint-on is used selectively for on-screen callouts and late additions.
What is the difference between CEA-608 and CEA-708? CEA-608 is the analog-era standard with fixed white-on-black styling and four channels. CEA-708 is the digital standard supporting fonts, colours, sizing, and positioning, and it can carry 608 data for backward compatibility.
Are subtitles and closed captions the same? No. Subtitles carry dialogue for viewers who can hear. Closed captions carry dialogue plus speaker identification and non-speech audio for viewers who cannot.
Do open captions satisfy accessibility requirements? They make text visible to everyone, but because viewers cannot toggle them, they do not replace closed captions for regulated broadcast delivery.
Where Digital Nirvana Fits
Once you know which types you need, the operational question becomes whether one workflow can produce all of them without duplicating effort.
TranceIQ handles transcription, captioning, and subtitle generation with the conformance rules each destination expects, so a single pass can output broadcast and streaming formats from the same source. Media Enrichment adds the managed human review layer that keeps accuracy and placement defensible at volume, including live captioning.
For teams whose caption work feeds a wider pipeline, Subs N Dubs extends the same source transcript into translated subtitles and dubbed audio, MonitorIQ validates caption presence and quality on the actual air feed, and MetadataIQ turns caption text into searchable metadata so the archive becomes discoverable as a by-product.
Expertise Behind the Standards
Caption types look like trivia until a delivery fails.
Knowing that 708 carries 608 inside it, that IMSC positioning behaves differently across OTT players, and that live placement rules matter as much as accuracy is the kind of knowledge that comes from operating inside broadcast delivery rather than reading a spec sheet. Digital Nirvana has built captioning workflows with broadcasters, OTT platforms, post-production houses, and education providers across exactly those constraints.
That is also why the human review layer is deliberate rather than decorative. Automated speech recognition handles volume. People handle proper nouns, speaker identification, and the placement calls that keep captions off a scoreboard. You can see how that combination works in practice across our customer success stories and our guide to automatic closed captioning software.
Conclusion
There is no single best type of closed caption. There is only the right combination for a specific piece of content going to a specific destination.
Decide the display style from whether the content is live. Decide the standard from whether it goes to broadcast or streaming. Decide the format from the destination’s spec sheet. Then build a workflow where one captioning pass feeds every output rather than starting over per platform.
Get those three decisions in the right order and the rework disappears.
Want the full picture on compliance obligations? Read our FCC caption rules guide for what the quality standards require across TV and streaming.
Key Takeaways
- Caption type is three separate decisions: display style, technical standard, and delivery format. Mixing them up causes most failed deliveries.
- Pop-on suits prerecorded content, roll-up is the live standard, and paint-on is a specialist option. Live or not live drives the choice.
- CEA-608 is the analog-era standard with fixed styling. CEA-708 is the digital standard with full styling control and backward compatibility.
- Subtitles assume the viewer can hear. Captions do not. A translated subtitle file is not an accessibility deliverable.
- SDH bridges the gap, carrying caption-level information through a subtitle-style delivery path used by streaming platforms.
- Broadcast needs SCC or MCC, OTT generally needs TTML or IMSC, and web needs WebVTT or SRT. One pass should feed all of them.
- The FCC’s four quality standards are accuracy, synchronicity, completeness, and placement, with a best-efforts allowance for live programming.
- Work backwards from your distribution list before commissioning captions, not forwards from production.
