What Is Closed Captioning? Definition, How It Works & When You Need It

Closed captioning (CC) is on-screen text that reproduces the dialogue and non-speech audio of a video — speaker labels, sound effects, music cues — designed so viewers who can't hear the audio can follow along. "Closed" means the captions can be toggled on or off by the viewer, unlike open captions which are burned into the video permanently. On TVs, streaming services, and YouTube, the CC button (usually a small [CC] icon) is what turns them on.

The name goes back to broadcast TV in the 1970s and 1980s. Closed captions are encoded as a separate data stream inside the video signal — for analog NTSC broadcasts, on line 21 of the vertical blanking interval; for digital ATSC broadcasts, in the CEA-708 caption data track. The viewer's TV decodes the caption data only when the CC feature is turned on. Open captions, by contrast, are burned into the video pixels themselves and are always visible. "Closed" is short for "closed to viewers who don't request them" — they're hidden by default.

Two subtle differences. (1) Purpose: closed captions assume the viewer cannot hear the audio and include speaker labels, sound effects, and music cues. Subtitles assume the viewer can hear and just need translation or reading assistance — typically dialogue only. (2) Regional terminology: in the US, "captions" usually means accessibility-grade tracks and "subtitles" means translation. In the UK and much of the EU, "subtitles" is used for both — BBC subtitles include accessibility content. For a deeper breakdown of the technical, legal, and regional distinctions, see captions vs subtitles.

SDH stands for Subtitles for the Deaf and Hard of Hearing. It combines the subtitle format (bottom-center text delivery, styled like translation subtitles) with caption-style content (speaker labels, sound effects, music cues in brackets). SDH is the standard accessibility format on Blu-ray, DVD, and streaming platforms (Netflix, Disney+, Amazon Prime Video) because their delivery pipelines don't carry broadcast caption data (CEA-608/708). When you see "English [CC]" or "English [SDH]" in a streaming platform's language menu, both are accessibility tracks — the platform may label them inconsistently.

Different accessibility feature for a different audience. Closed captions describe the audio track for viewers who can't hear. Audio descriptions describe the visual content for viewers who can't see — a narrator voice track that explains on-screen action during pauses in dialogue. Netflix, streaming platforms, and broadcast increasingly provide both as separate tracks. The US CVAA (21st Century Communications and Video Accessibility Act) requires audio descriptions on major broadcast channels for a set number of hours per week; the EU EAA requires member states to phase in similar requirements.

Four main pipelines in 2026. (1) Broadcast: human transcriptionists produce captions in real time or near real time, encoded as CEA-608/708 in the video signal. Rev, VITAC, and 3PlayMedia dominate this. (2) Streaming/OTT (Netflix, Disney+, HBO Max): captions delivered as separate .srt / .vtt / TTML tracks, typically produced by human-plus-AI review workflows. (3) YouTube auto-captions: Google's ML model generates captions automatically at roughly 75-85% English accuracy, lower for accented speech and non-English languages. (4) AI transcription tools (VexaScribe, Otter, Descript, Rev AI, Deepgram): Whisper Large-v3 or comparable models produce transcripts at 90-95% accuracy on clean English, exportable as SRT or VTT for attaching to your own videos.

Depends on the audience and platform. US broadcast TV and cable: yes, required by FCC rules 47 CFR §79.1. US online video previously aired on TV with captions: yes, required by the CVAA. Educational, government, and large-organization websites: strongly recommended under WCAG 2.1 Level A (Success Criterion 1.2.2 requires captions on prerecorded video) and often required under ADA case law (NAD v. Netflix 2012, NAD v. Harvard and MIT 2019-2020). EU services and e-commerce with video content: European Accessibility Act (Directive 2019/882) took effect 28 June 2025 with phased compliance. UK broadcast: Ofcom Code sets subtitling targets. Personal social media videos: not legally required, but 50-80% of viewers watch with audio off, so captions materially affect engagement.

No — a common Ofcom study reported that roughly 80% of UK caption users are not deaf or hard-of-hearing. Modern caption users include non-native speakers improving comprehension, viewers in noisy environments (gyms, public transport, open offices), viewers watching muted (social media autoplay), users with attention or processing differences, learners studying with media, and viewers of accented or technical content who catch details more reliably by reading. Netflix, YouTube, and Amazon Prime report caption usage rates of 50-80% among general audiences.

Several, each for different delivery contexts. SRT (SubRip) is the most universal — plain text with sequence numbers and timestamps; supported by YouTube, Vimeo, Premiere, DaVinci, CapCut, VLC, and virtually every platform. WebVTT (.vtt) is the W3C standard for HTML5 web video, required for the HTML5 <track> element; supports positioning, colors, regions, and metadata. SCC (Scenarist Closed Caption) is the binary representation of CEA-608 data used in US broadcast. TTML and IMSC are XML-based formats used in OTT delivery (Netflix uses IMSC1.1). ASS/SSA supports advanced positioning, popular in anime and fan subtitles. For most creator workflows, SRT is the right default.