Reading Speed in Captions: What the Characters-Per-Second Rule Means

Characters per second, often shortened to CPS, is a simple count. Take the number of letters, spaces and punctuation marks in a caption, then divide by the number of seconds the caption stays on screen. A caption with 42 characters shown for two seconds reads at 21 CPS. That single figure tells you how hard a viewer has to work to keep up with the words.

The reason the term matters is that speech and reading do not move at the same pace. People often speak in bursts, then pause, then speak again. A viewer reading captions cannot see the next line until it appears, so a number that looks fine on paper may still feel rushed on screen.

When I teach this module, I ask learners to count one caption by hand before they touch any software. It makes the rule feel less abstract. Once you have counted a dozen lines, you start to sense when a caption is going to be too dense.

• Count letters, spaces and punctuation together
• Divide by on-screen seconds, not by speaking time
• A low number is easier to read, a high number is harder

Most teams settle on a range rather than one exact number, because the right pace depends on who is watching and what is being said. Recorded interviews and talks for a general audience usually sit at the gentler end. Viewers who are new to reading captions, or who are watching on a small phone screen, need that extra room.

Technical vocabulary pushes the number down further. A word like a long product name or a place name takes longer to read than its character count suggests, because the eye stops on unfamiliar shapes. If a speaker lists several terms in a row, give each caption more time even when the audio is quick.

I tell learners to treat the range as a guide for decisions, not a score to chase. A caption that sits slightly over the range for half a second is rarely a problem. A caption that is far over it for several seconds is worth rewriting.

Reading speed is only half of the picture. The other half is how the text is split across lines. A caption of two short lines is easier to take in than one long line that runs across the whole screen, and a break in the wrong place can slow a reader even when the CPS number looks safe.

Break at natural pauses in the sentence: after a comma, before a conjunction, or between a subject and its verb phrase. Avoid leaving a single short word, such as an article or a preposition, stranded at the end of the first line. Try to keep the two lines roughly balanced so the eye does not have to travel far.

If a sentence is simply too long for two lines, split it into two separate captions and give each its own timing. This is usually kinder to the viewer than squeezing the words smaller or faster.

• Split at commas, conjunctions and clause boundaries
• Keep names and titles together on one line
• Prefer two balanced lines over one long line

A caption should appear just as the speaker begins the words it carries, and leave shortly after they finish. Appearing a fraction early can be acceptable, since viewers glance at text a moment before they register the voice. Appearing late is worse, because the audio gets ahead of the words and the whole experience feels out of step.

Leave a short hold after the last word so the viewer can finish reading. When two captions follow each other with no gap, they can blur together, so a small space between them helps the eye reset. When the speaker pauses for a long moment, clear the screen rather than leaving old text hanging.

In practice, I place the start time first against the waveform and the picture, then adjust the end time until the number feels right. Playing the clip back at normal speed, with the sound on, is the check that catches most timing slips.

Some speakers talk very fast, and a literal rendering of every word would push the reading speed far past the comfortable range. In those cases you have a choice. You can lightly condense the wording, dropping filler such as repeated starts or phrases that add nothing, while keeping the meaning and the speaker's voice intact.

Condensing should be a last step, not a habit. Interviews where exact wording matters, such as a quoted statement or a technical explanation, are better handled by giving the text more time, even if that means overlapping slightly into the next pause.

Whatever you decide, apply it consistently across a project. A viewer adjusts to a steady style quickly, but notices when one section is trimmed hard and another is word for word. Write down your choice in a short style note so that later captions follow the same approach.

• Trim fillers before trimming meaning
• Keep quoted or technical wording exact
• Record your decision in a style note

The most useful practice I can suggest is to review a finished caption file with the picture muted and only the captions visible, then do the opposite with the sound on. The first pass shows whether the text is readable on its own. The second shows whether the timing feels natural against the voice.

As the autumn weeks bring more event and talk recordings to prepare, this quick double check is a good way to catch uneven pacing before delivery. Note any caption that made you pause or reread, then adjust its line break or hold time.

Over several projects you will begin to feel the right pace without counting. The number remains a helpful anchor, but your own judgement, trained by repeated review, does most of the work.