Closed captions vs subtitles: they are not the same thing

The one-line difference (and the test)

W3C's HTML5 spec draws the line precisely: subtitles are a "transcription or translation of the dialogue when sound is available but not understood," while captions are a "transcription or translation of the dialogue, sound effects, relevant musical cues, and other relevant audio information" for when sound is unavailable or the viewer is deaf or hard of hearing. Same screen position, different job.

The two-second test: mute the video. If the text track still tells a complete story — who is speaking, that music swells, that a door slammed off-screen — it is a caption track. If it only carries the words people say, it is a subtitle track. A deaf viewer watching caption-less subtitles misses a phone ringing, an alarm, or sarcasm carried entirely by tone — information a hearing viewer never has to think about.

The file extension does not decide anything: the same SRT file is a caption track or a subtitle track depending on what the text contains. YouTube muddies this further by labeling everything "Subtitles/CC" in its interface — the name of the track in Studio tells you nothing about what the file actually carries.

SDH (Subtitles for the Deaf and hard of hearing) is the bridge category: caption-style information delivered in a subtitle file format. Streaming platforms and DVDs use it because their subtitle pipelines handle the formatting; the content inside is caption content. If a deliverable spec asks for SDH, it wants speaker IDs and bracketed sound cues like [tires screeching] inside the subtitle file.

What the law requires in 2026

In the US, closed captions are a legal requirement for most public-facing video — subtitles are not, because translation serves convenience, not access. The four layers that matter:

FCC (broadcast & previously-aired online video). Closed captioning is required for virtually all new English and Spanish programming on broadcast and cable TV. Quality is enforced on four pillars under 47 CFR § 79.1(j): accuracy, synchronicity, completeness, and proper placement. The CVAA (2010) extends the requirement to internet video that first aired on US television with captions — if your clip aired on TV, the web copy needs captions too.

ADA (public-facing business video). Courts have interpreted ADA Title III to cover websites, and video content on them, as places of public accommodation. The DOJ's April 2024 final rule formally adopts WCAG 2.1 Level AA as the technical standard for state and local government content (Title II), with the first compliance deadline for larger entities passing on April 24, 2026. Title III obligations for private businesses come from court decisions and settlements rather than one codified rule, but the practical bar is the same: synchronized captions on prerecorded video.

WCAG 2.1 (the technical standard courts cite). Success Criterion 1.2.2 (Level A) requires captions for all prerecorded synchronized media with audio — no exemption for small businesses, embeds, or "just a quick clip." Success Criterion 1.2.4 (Level AA) adds live video: webinars and live streams need captions too. A separate transcript page does not satisfy 1.2.2; the text must be time-synced inside the player, which means an SRT/VTT track, not a linked document.

Open captions do not satisfy accessibility law. Burned-in text is visible to everyone, but compliance standards require a separate, toggleable track the viewer can switch off. Burned-in captions are for social feeds and kiosks where no player toggle exists — fine for reach, not a substitute for a closed track on anything that must meet ADA, FCC, or WCAG.

Outside the US: the European Accessibility Act requires captions for digital services across the EU, and Section 508 keeps the same WCAG-derived bar for US federal agencies and contractors. There is no legally mandated accuracy percentage anywhere — the widely-cited "99%" figure comes from private accessibility settlements and industry practice, not a statute. Platform auto-captions are a starting point, not a compliance deliverable: WCAG requires accurate captions, and auto-generated tracks routinely fail on names, technical terms, and speaker IDs.

Which one to upload, by platform

The platform decides the format; the audience decides the content. One corrected transcript produces both deliverables — captions keep the non-speech cues, translations strip them.

The creator default in 2026 is closed captions on everything long-form (YouTube, courses, webinars — where the player supports a toggle) and open captions on short-form (TikTok, Reels, Shorts — where autoplay-mute means most viewers never touch a CC button and the text must be guaranteed). Roughly half of US adults report using captions at least some of the time, and 15% report some degree of hearing loss — captions are the rare accessibility feature that serves everyone else too.

Subtitles earn their place when the audience crosses a language line: international releases, language learning, and local-market reach. A truly global release ships both — closed captions in the source language for accessibility, subtitle tracks in target languages for translation. Both come from the same corrected transcript, so the marginal cost of the second deliverable is formatting, not transcription.

Build both from one transcript, locally

Every track above starts the same way: an accurate, timestamped transcript. VocalFuse produces it locally on Windows with a Whisper-class engine — audio or video in, punctuated cues out, SRT/VTT/TXT export from the same pass, and nothing ever uploaded. No per-minute meter because no cloud API is listening: the free tier covers transcription, Basic ($5/mo) adds dictation, and Pro ($10/mo) adds AI notes and summaries.

From that transcript: add bracketed sound cues and speaker labels for the caption track ([music swells], SARAH:), strip them for the subtitle track, and keep timestamps identical in both. Then follow the destination table — SRT to YouTube, VTT to your web player, burned-in text for the vertical feeds. If you need the file format details first, the SRT vs VTT decision guide covers which format each destination accepts, and the SRT creation guide covers the block format itself.

Cross-border content follows the same pipeline one step later: the SRT translation guide covers keeping timestamps intact while the text changes language. For meetings and calls that later become video, meeting transcription and free voice transcription feed the same corrected-transcript workflow.

Frequently asked questions