What are karaoke captions?

Ordinary subtitles show a sentence at a time. Karaoke captions need a time on every word: speech recognition gives each word a start and an end, the words are grouped into short lines (3 at a time in Clyper's Karaoke style), and each one changes color as it's said.

The name comes from karaoke subtitles, and subtitle files still mark them up that way. In the Advanced SubStation Alpha (ASS) format, a \kf tag before each word sweeps its fill from one color to another, left to right, over that word's duration, timed in hundredths of a second. Clyper writes its captions as ASS, and its Karaoke style uses exactly that tag. For right-to-left scripts like Arabic and Hebrew, each word fills at once instead, since a left-to-right sweep would run backwards through the word.

In the wider sense, any caption that marks the spoken word gets called karaoke style: the word changing color (Pop), a highlight behind it (Highlight) or one big word at a time (One word). Here are all 5 of Clyper's:

Plenty of people read along rather than listen, and not only people who can't hear. The BBC's Subtitle Guidelines say around 10% of broadcast viewers use subtitles regularly, rising to 35% for some online content, and that most of them aren't hard of hearing.

A highlight does two things a plain line can't:

• It shows where the voice is, so someone watching on mute reads at the speaker's pace instead of racing ahead and waiting for the next line.
• It moves. The color stepping from word to word gives the eye something to follow, which suits a feed where people decide in a second whether to stay.

Karaoke captions are burned into the picture: the font, the colors and the fill are part of the clip, so they look the same on every app. Whether a clip needs captions or subtitles in the first place is in captions vs subtitles.

• Two colors. One for the words and one accent for the spoken word or the fill. With more, nothing stands out.
• A heavy font with an outline, big enough to read at arm's length on a small phone.
• A few words at a time. 3 on a line is plenty; for fast talkers, One word shows one at a time.
• Exact timing. A highlight that lags behind the voice looks broken, so the times have to come from the audio itself. Clyper times each clip's words from that clip's own audio.
• The right words. A highlight draws the eye to every word, a misheard name included. Fix them before you post.
• Clear of the buttons. Below the face and above the bottom of the frame, inside the safe zone.

To try the styles on a video you already have, the free add captions to video tool burns them into a vertical video of up to 5 minutes, with no watermark. Sizes, fonts and placement in more depth are in how to caption short videos.