CC vs subtitles vs forced narrative
Closed captions include non-speech cues. Subtitles are dialogue. Forced narrative shows only the lines you need with sound on.
These three labels are different files and different jobs. Mixing them breaks accessibility and platform uploads.
Creators and editors often treat "CC," "subs," and "forced" as synonyms for "words on screen." Platforms do not. Delivery specs do not. A mute TikTok burn-in is not a Netflix forced narrative track, and a YouTube auto CC file is not a localization subtitle. If you ship the wrong track type, you get missing cues for deaf viewers, duplicate dialogue over clear audio, or a soft-sub upload that platforms ignore because style lived in the wrong container.
This page separates the three jobs, shows when burn-in is the right container, and points you at the Versely paths that keep short-form captions honest without pretending every export is an accessibility deliverable.
Closed captions: speech plus the room
Closed captions carry dialogue and non-speech information. Door knocks, music stings, speaker IDs, laughter, "[phone rings]," "[soft piano]." The audience for CC includes people who cannot hear the mix, or who need the mix named because hearing alone is not enough. That is why CC files are denser than dialogue-only subs, and why stripping sound cues to "clean up" a CC track is not a style choice. It is a cut to the accessibility job.
On broadcast and many streaming pipelines, closed captions arrive as a toggleable track (CEA-608/708, timed text profiles, platform CC). The viewer can turn them on. The picture does not permanently include the glyphs. That is the opposite of a burned-in short where mute scrollers never get a toggle.
For short-form social, most creators never ship a true CC sidecar. They burn readable type into pixels because TikTok, Reels, and Shorts autoplay muted and do not guarantee your styled track will appear in the first second. That burn can still contain CC-quality information (speaker labels, sound cues) if you write them. Most teams do not. They transcribe speech only. Call that what it is: speech captions for mute retention, not a full CC deliverable for compliance. If a client asks for WCAG-minded captions on a long cut, do not hand them a karaoke Snap export and call it closed captions.
When you do need a sidecar for review or archive, keep style out of the wrong place. Our note on burned-in captions vs SRT and VTT sidecars covers when pixels win and when a text track wins. Sidecars are editable and accessible to players that support them. Burns are what mute feeds actually show.
Practical test: if a deaf editor on your team would miss plot or tone without a cue that is not spoken dialogue, the track still has CC work left. If the only job is "mute scroller can read the hook," you are in short-form caption territory. Same letters on screen. Different brief.
Subtitles: dialogue for sound-on (or translation)
Subtitles, in the classic sense, are dialogue (and sometimes song lyrics) timed for viewers who can hear ambient sound but need the words: another language, noisy room, accent clarity. Full subtitle files aim to cover the speech track end to end. They usually omit the rich non-speech layer that CC includes, because the mix is assumed to be available.
Localization teams live here. You lock picture, lock an English dialogue list, then produce language tracks. Forcing every language into a burn-in is expensive and locks you out of late copy fixes. Soft subs (SRT, VTT, TTML, IMSC depending on the platform) stay editable. Soft subs also fail the mute-feed retention job on social, because the platform may delay, restyle, or hide them until after the hook.
Creators blur "subtitles" with "captions" because CapCut and phone editors label every timed line a subtitle. Internally, say what you mean: dialogue-only timed text, or speech-plus-cues, or forced-only. Upload forms that ask "subtitle file" often expect dialogue. Upload forms that ask for "captions" may expect CC semantics. Guessing wrong is how a film festival rejects your DCP package or how YouTube shows a sparse track that never names the score.
For Versely short-form work, add captions to a video is the speech-to-styled-burn path. It is not a localization desk. If you need a Spanish full subtitle file for a 12-minute YouTube, that is a different workflow: export a clean transcript, translate, retime, deliver soft subs. Do not burn Snap karaoke into a bilingual master and hope the platform "figures it out."
Forced narrative: only the lines you cannot skip
Forced narrative (often "forced subs" or "forced narrative track") shows selected lines even when the viewer has full-language audio on and other subs off. Classic uses: a shop sign in a foreign language the character reads aloud in English later; an alien language the audience must understand; a plot-critical text message on a phone screen that the mix never speaks. The point is scarcity. Forced narrative is not a full subtitle pass with most cues deleted in the UI. It is an authored set of must-see lines.
Mixing forced narrative with a full subtitle track is a packaging problem. Players that support forced flags will show the forced events when "subs off" is selected. If you bake forced lines into the same full track without flags, viewers who want "subs off" still see nothing, or you burn forced lines into picture so everyone sees them always. Cinema and premium streaming care about this. A 15-second UGC ad almost never needs a forced narrative track in the SMPTE sense. It may still need a burned line for a price or a foreign logo the voice never says.
We already compared forced narrative subs versus full subtitle files. The retention lesson for social is narrower: if a line must be read with sound on (a disclaimer, a SKU, a foreign packaging claim), author it as overlay or burn it. Do not rely on platform auto CC to invent a forced event. Auto CC listens for speech. It will not caption a silent price tag.
When agencies say "add forced subs" on a social brief, ask which lines, in which language, and whether burn-in is acceptable. Nine times out of ten they mean "make sure this one sentence is readable." That is a one-line burn or overlay, not a forced narrative package.
What to ship for short-form vs long-form
Short-form mute feeds (TikTok, Reels, Shorts, many Stories): burn styled speech captions into the file before upload when retention is the job. Prefer a caption family that matches the cut. Keep a sidecar SRT only if you need editability later. Do not call that burn "closed captions" in a client accessibility report unless you actually authored non-speech cues and tested with that audience.
Long-form sound-on (YouTube education, podcast video, webinars): soft captions or subs usually win. Viewers toggle. You can fix typos without re-exporting a 40-minute mezzanine. Platform auto CC is a draft, not a delivery. If the channel cares about deaf access, budget a CC pass, not only a speech transcript.
Hybrid (trailers, multilingual ads): forced narrative or burned key lines for plot-critical text, plus a full soft track for the language the buyer selected. Keep the jobs in separate files when the platform allows. When it does not, burn the forced lines and attach the full track as soft.
Naming discipline saves rework. In your shot list or CAP file notes, write CC, dialogue subs, or forced, not "text." In Versely, generate speech captions for the mute job, then decide burn vs sidecar with the burned-in vs SRT guide linked above. If a producer asks for all three labels on one social cut, push back with this page. One file can hold burned speech for the feed. It cannot honestly be a full CC package, a localization subtitle set, and a forced narrative track at once.
Credits on Versely caption renders are usage credits, not a currency price tag in this article. Preview before you commit a dynamic look. Wrong track type wastes credits and calendar time more often than wrong kerning does.
Ready to burn speech captions for mute feeds without confusing them for a festival CC package? Create a Versely account and run the caption pass on a short cut first.
FAQ
Are closed captions the same as subtitles?
No. Closed captions include non-speech cues and speaker information for viewers who need more than dialogue. Subtitles are primarily dialogue (often for translation or clarity) and usually assume the soundtrack is available. Short-form "captions" in creator tools are often speech-only burns, which match neither label perfectly.
Do I need forced narrative on TikTok?
Almost never as a formal track. If one line must appear with sound on (a legal line, a foreign sign, a price the VO skips), burn or overlay that line. Do not expect platform auto captions to invent it from silence.
Should I upload SRT or burn captions for Reels?
Burn when style, timing, and the first-second mute hook matter. Keep an SRT for archive or for platforms that honor soft subs on long-form. For mute autoplay, pixels you control beat a late platform restyle.