Timing and Sync

How Long Should a Subtitle Stay on Screen

Minimum and maximum display durations for subtitles, why guides differ, and the trade-offs between reading time, speech timing and the picture.

On this page
  1. The question behind the question
  2. The lower limit
  3. The upper limit
  4. Matching duration to the amount of text
  5. Extending past the end of speech
  6. The audience matters
  7. The picture matters too
  8. Gaps take time too
  9. Checking in practice
  10. A sensible default when there is no specification

The question behind the question

Display duration is the length of time one subtitle is visible, from its start time to its end time. People often ask for a single number. The honest answer is that there is a range, and the right place in that range depends on how much text the subtitle holds, who is reading it and what is happening in the video.

The lower limit

Very short subtitles are hard to read. A cue that appears for less than about a second can seem to flash, and the viewer may only register that something appeared. Many style guides set a minimum of roughly one second, and some allow slightly less for a single short word. These are typical figures, not universal rules, and individual specifications differ.

The minimum is usually applied even when the speech is shorter. If a person says “Yes.” in a quarter of a second, the cue still often stays long enough to be read. That works only if the next speech does not start too soon, which is why short utterances are sometimes merged with neighbors.

The upper limit

There is also a ceiling. A cue that stays for too long can look stuck. If the speaker has stopped, the viewer who finished reading may wonder whether something is wrong. A long-lasting cue also covers a part of the picture for longer. Many guides put the maximum for a normal two-line cue at about six or seven seconds. Beyond that, the text is often split into two cues, or the second half is shown separately.

Again, these are common conventions and not fixed laws. Material with long, slow speech, or with onscreen text, may be handled differently.

Matching duration to the amount of text

A fixed duration for every cue would not work, because cues vary in size. A short phrase needs less time than two full lines. This is why display duration is tied to reading speed. Reading speed expresses how many characters or words a viewer is expected to read per second. If a cue has 70 characters and the guideline allows 17 characters per second, the cue needs at least about 4.1 seconds, since 70 divided by 17 is roughly 4.12. That is a hypothetical example to show the arithmetic. Real guidelines differ, and the right rate depends on the audience and language.

When the speech is too short for the text, the writer has choices:

  • Shorten the text, using text condensation to keep the meaning with fewer words.
  • Start the cue slightly before the speech or end it slightly after, if neighboring speech allows.
  • Split the text across two cues, if the speech has a natural break.

Each choice has a cost. Condensing loses detail, extending can leave the text on screen when the speaker has stopped, and splitting adds more changes for the eye to follow.

Extending past the end of speech

Many subtitlers let a cue stay on screen a little after the last word is spoken. This gives slower readers time to finish. The extension must not collide with the next cue or lead to an overlap. It is also usually kept modest, since a cue that lingers too long after the voice has stopped looks out of step.

The audience matters

Viewers who are less experienced readers, including younger children and people learning a language, typically need more time per word. Older viewers may also appreciate longer display, while some experienced readers prefer text that keeps pace closely. A subtitle for a young audience might therefore have fewer characters per cue or a lower reading speed target, which makes it last longer. Guidance on these groups differs, so follow the specification for your project.

The picture matters too

Duration is not only about reading. The viewer also wants to see the picture. Fast action, on-screen text, charts or graphics compete with subtitles for attention. A cue that stays long over a busy image can be demanding. Where the picture changes with a cut, editors often prefer cues to end at or near the cut, which can force a shorter or longer duration than the speech alone suggests. That is another reason durations cannot be set by a formula.

Gaps take time too

The blank interval between consecutive cues comes out of the available time. A cue that ends a few frames early to leave a gap loses that time. When speech is fast, this can push a cue below its comfortable duration. Writers balance the two, and may relax one rule slightly to satisfy the other.

Checking in practice

Tools can flag cues that are shorter than the minimum or longer than the maximum, and cues whose text is too long for their duration. These flags are useful starting points for subtitle proofreading. The final judgment still needs someone to watch the video at normal speed and ask whether each cue can be read comfortably without rewinding. If viewers are likely to pause or rewind to finish a line, it is probably too short.

A sensible default when there is no specification

If you have no guidelines to follow, a conservative approach is to keep cues at least about a second long, avoid going much beyond six or seven seconds, and let the amount of text set the time within that range. Then adjust for the audience, the pace of the speech and the picture. A small amount of testing with real viewers, where possible, will tell you more than any single number.

The complete guide to timing and sync

How Subtitle Timing Works
Read the guide

Keep exploring

Guides from other topics

All topics →