Subtitles and Captions for Accessibility and Reach

It's easy to think of captions purely as a legal or accessibility checkbox, but the numbers tell a different story: a large share of video views on social platforms happen with the sound off, and captioned videos are consistently watched for longer than uncaptioned ones, regardless of whether the viewer is deaf, hard of hearing, in a quiet office, or simply scrolling on a bus. Captions have quietly become a core part of the viewing experience for almost every audience, not a niche accommodation.

Subtitles assume the viewer can hear the audio and simply need the dialogue translated or transcribed, typically for a different language. Captions are written for viewers who can't hear the audio at all, so they also describe relevant non-speech sound — [tense music], [door slams], [laughter] — because that information carries meaning the visuals alone don't convey. Choosing the wrong one for your audience genuinely changes what needs to be included.

The two-line, two-second rule

Most accessibility guidelines recommend keeping captions to no more than two lines on screen at once, and displaying each caption block for at least one to two seconds — long enough to be read comfortably, even by a slower reader, without lingering so long it falls out of sync with speech.

Reading speed limits

A commonly used benchmark caps caption reading speed around 15-17 characters per second, or roughly 150-180 words per minute — noticeably slower than typical spoken pace, which is why long, fast dialogue sometimes needs to be lightly condensed rather than transcribed word for word.

Perfectly verbatim captions aren't always the clearest captions. Filler words ('um', 'you know', false starts) can usually be trimmed without changing meaning, and trimming them often makes a caption easier to read in the limited time it's on screen. The goal is fidelity to meaning, not a court-transcript level of literal accuracy.

Break caption lines at natural grammatical pauses — after a clause, not mid-phrase — so each line reads as a coherent chunk rather than an arbitrary character-count cutoff.

Use a high-contrast background or outline behind caption text so it stays readable over both light and dark footage. Keep font size large enough to read on a small phone screen — what looks fine on a desktop preview can be nearly illegible at mobile size. Position captions to avoid covering on-screen text, faces, or other important visual elements.

Burned-in (open) captions are rendered directly into the video file and always display, which is the safer choice for social platforms where many viewers won't manually enable captions. Caption files (like SRT or VTT) attached to a video give viewers the choice to toggle captions on or off and support multiple languages from a single video file — the better choice for platforms like YouTube where captions are optional and searchable.

Caption and subtitle files are plain text, which means the spoken content of your video becomes crawlable and indexable in a way raw audio and video pixels never are. Platforms and search engines use caption transcripts to understand what a video is about, match it to search queries, and sometimes surface specific timestamped moments directly in search results.

Generate an initial automatic transcript. Correct misheard words, especially names, brand terms and technical vocabulary the auto-generator won't know. Trim filler words and break lines at natural clause boundaries. Check each caption block against the two-line and reading-speed guidelines. Choose burned-in captions for short-form social clips and a toggleable caption file for longer platform-native uploads.

Captions improve watch time and reach for all viewers, not just those who are deaf or hard of hearing.

Keep captions to two lines, readable for at least a second or two, and paced under roughly 150-180 words per minute of reading speed.

Choose burned-in captions for sound-off social platforms and toggleable caption files for searchable, multilingual platform uploads.