01 / Start with access
Give viewers another route into the audio.
Captions are synchronized text alternatives for spoken dialogue and meaningful non-speech audio. They support people who are Deaf or hard of hearing and people watching in a sound-sensitive place or with their sound turned off.
Closed captions can be switched on or off in a supporting player. Open captions are burned into the picture and stay visible. Closed captions give viewers control; open captions travel with the exported video but cannot be toggled.
The research snapshot supplied for this guide cites an approximate 85% muted-viewing estimate. Its scope and method were not established here, so treat it as a directional estimate, not a universal or current viewing rate.
Sources: W3C WCAG 2.2: Captions (Prerecorded), SC 1.2.2Section 508: Captions and Transcripts
02 / Keep meaning intact
Caption more than the words.
A viewer should be able to follow the audio without having to guess who spoke or which sounds changed the meaning.
Dialogue, accurately timed
Keep the wording faithful to what was said and synchronize each caption with the corresponding speech or sound. Use punctuation that makes the thought easy to follow.
Speaker turns
Identify a speaker by name or role when it is not clear from the picture who is talking, especially when the speaker is off-screen or the conversation changes hands.
Meaningful non-speech audio
Include sounds such as laughter, an alarm, or a significant music change when a viewer needs that information to understand the moment. A brief cue like “[laughter]” can be enough.
Important visual-only information
Captions describe audio, not images. If an action, chart, or on-screen statement carries essential information that the audio does not convey, provide a separate description or text alternative.
Sources: W3C WCAG 2.2: Captions (Prerecorded), SC 1.2.2W3C WCAG 2.2: Audio Description or Media Alternative, SC 1.2.3Section 508: Captions and Transcripts
03 / Make it easy to read
Use phone-first caption settings as a starting point.
These are practical starting points from the supplied research and Section 508 captioning guidance. Adjust them for the player, language, and pace of the video.
Type and size
Choose a sans-serif face such as Helvetica or Arial. Section 508 captioning guidance gives 18 points as a default size; increase or adjust it when the player or screen calls for it.
Line length
Keep captions to no more than two lines. Aim for about 42–45 characters per line; the shorter end can leave more room on a narrow phone screen.
Reading pace
The supplied research snapshot suggests roughly 17–20 characters per second as a readability target. Treat that range as editorial guidance, not a WCAG or universal legal threshold; split dense wording or let it stay on screen longer when possible.
04 / Protect the picture
Leave important visuals unobstructed.
A centered position in the lower third is a useful default, but move captions when they cover faces, hands, product labels, signs, or app interface details. Keep other on-screen text in the upper two-thirds where possible, since many players place captions low in the frame.
Check the actual crop and player controls: the safest position can change between a vertical feed, a preview, and the exported file.
Sources: Section 508: Captions and TranscriptsW3C WCAG 2.2: Captions (Prerecorded), SC 1.2.2
05 / Keep contrast steady
Make the text hold up over changing footage.
White text on a dark translucent backing is a practical starting style, not a required look. Keep the treatment consistent and check it over both bright and dark frames; a backing that looks strong on one frame may fade into another.
WCAG 2.2 Success Criterion 1.4.3 specifies at least 4.5:1 contrast for normal text and 3:1 for large-scale text, within the criterion’s scope and exceptions. Use those ratios as contrast targets where the criterion applies; the backing color itself is a styling choice, not proof that the ratio is met.
06 / Before you post
Review the whole clip in context.
Auto-generated captions are a draft. Run through this short checklist before publishing:
Correct the words
Check names, terminology, spelling, punctuation, and speaker turns against the recording.
Check timing and pace
Watch captions with audio and confirm they arrive with the speech, stay long enough to read, and do not flash by in dense blocks.
Watch it muted
Confirm a viewer can follow the spoken ideas without sound, including meaningful non-speech audio cues.
Check changing footage
Look for weak contrast and make sure captions do not hide a face, hand, label, interface detail, or other important visual.
Preview the export and player
Review the exported video in its intended player. Confirm open captions remain legible, or that closed captions can be enabled and render as expected.
Sources: Section 508: Captions and TranscriptsW3C WCAG 2.2: Captions (Prerecorded), SC 1.2.2
07 / Platform notes
Use each platform’s documented caption path.
These are platform-specific instructions from official help pages. They describe the documented features for each service, not a rule that applies to every platform.
LinkedIn
LinkedIn Help documents auto-generated captions for uploaded videos, with options to review them before viewers see them and edit them after generation. A separate LinkedIn guide describes attaching an SRT file for closed captions on desktop.
X
X Help documents uploading an SRT caption file for a video post and currently says this feature is available on web only. Check X’s help page for the latest interface and availability; this is not a universal X rule for every way of posting video.
Sources: LinkedIn Help: Auto captions for videosLinkedIn Help: Add closed captions to videosX Help Center: Upload a caption (.srt) file to Posts
08 / What the standards say
Keep WCAG levels separate from platform policy.
Under WCAG 2.2, SC 1.2.2, Captions (Prerecorded), is Level A. SC 1.2.4, Captions (Live), is Level AA when the respective criteria apply. These are WCAG conformance levels, not LinkedIn policy. Section 508’s captioning page is also a useful source of authoring guidance; the standards or laws that apply to a specific organization depend on its context.
The supplied research snapshot also proposes at least 99% transcription accuracy and synchronization within 100 milliseconds. Treat those figures only as report-supplied precision targets: the W3C and Section 508 sources consulted here do not establish them as universal legal or WCAG thresholds.
Sources: W3C WCAG 2.2: Captions (Prerecorded), SC 1.2.2W3C WCAG 2.2: Captions (Live), SC 1.2.4W3C WCAG 2.2: Audio Description or Media Alternative, SC 1.2.3Section 508: Captions and Transcripts