Speech is faster than comfortable reading, especially on ads and listicles. A captioner that times 1:1 with a fast talker will flash lines the eye cannot finish. Broadcast practice has long used a band around 160–180 words per minute for adult same-language captions, with slower targets for children's content and for a second language. Those numbers are guidelines, not a switch in the file format. What the viewer actually gets is cue duration versus word count.
The failure mode is specific: the first two words land, the rest vanish, the next cue is already there. Retention software will not tell you this. You only see it by watching the cut on mute at phone size. Word-highlight (karaoke) families can feel faster than they are because the accent moves; they still have to leave the line up long enough to finish. A hook line set as an overlay is not exempt — if it is the claim, it has to survive the first two seconds, not decorate them.
Fixing reading speed is almost never 'pick a punchier preset.' It is splitting the cue, cutting the sentence, or holding the last line a beat after the mouth stops. Versely's reading-speed checker flags cues against a WPM cap you set; the caption generator will not slow a speaker down for you.
In practice
- Watch the captioned cut on mute. If you cannot finish a line, the WPM is too high, regardless of what a desktop preview suggested.
- Split at a clause rather than speeding the next cue to 'keep up' with the voice.
- Hold the last caption a beat after the sentence ends; matching the mouth exactly is how the last word gets clipped.
The mistake to avoid
Matching caption in/out to the waveform and calling it done. The ear can follow a rate the eye cannot.
Where you will run into it
- AI Caption Generator — Drop a video, get timed captions. Free, on your device, no upload.
Related terms
Caption line length
Caption line length is how many characters sit on one cue. Latin-script lines are commonly kept near 42 characters so a phone can still read them.
Caption timing
Caption timing is when each line appears and disappears relative to the speech — the clock, not the words and not the look.
Burned-in captions
Burned-in captions meaning: subtitles rendered into pixels, so viewers cannot turn them off or lose them on re-upload. Closed captions are a separate track.
Hook rate
Hook rate meaning: the share of people shown a video who are still watching a few seconds in. It grades the opening, not the body.
Retention curve
Retention curve meaning: how many starters are still watching at each second. Watch time is the area under that curve.
The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync, in your browser or on your phone.
Free on the web and on iPhone, one account.