What timing means for subtitles
Every subtitle has two jobs. It must say the right thing, and it must say it at the right moment. Timing is the second job. It covers when a piece of text appears, when it disappears, how long it stays, and how it lines up with the voice and the picture.
Viewers rarely notice good timing. They notice bad timing immediately: text that arrives after the speaker has finished, lines that vanish before they can be read, or a caption that flickers by too quickly to take in. Because reading takes effort, timing decides whether subtitles feel like a help or a distraction.
The basic unit
Subtitles are stored as a list of short timed pieces of text. Each one has a start time, an end time and the words to show between them. These pieces are usually called cues. Cutting a script into cues, and choosing the start and end of each one, is the heart of timing work.
Those times are written as timecodes, which are positions measured from the start of the video. A subtitle file is, at its simplest, a list of timecode pairs with text attached. Nothing in the file “listens” to the audio. The text appears only because a player compares the current playback position with the times in the file.
When a cue should start
A common starting point is to show text at about the moment the speaker begins talking. If the text appears much earlier, it gives away what is about to be said. If it appears late, the viewer hears words they cannot yet read and may lose the thread.
Subtitlers often make small adjustments to this rule. Because the eye needs a moment to find new text, some people start a cue a few frames before the speech begins, while others stay close to the exact start of the voice. Practice varies between teams and between style guides, so there is no single correct offset.
When a cue should end
Ending a cue is a balance. The text should stay long enough to be read comfortably, yet it should not hang around after the speaker has moved on. Many guides allow a cue to linger for a short time after the speech ends, which gives slower readers a chance to finish. When another speaker begins immediately, the old cue has to end to make way for the new one.
The length of time a cue is visible is its display duration, and the trade-offs are real. Very short cues are hard to read and very long cues can feel stale or can cover the picture for too long.
Reading speed sets the limits
People read at different speeds, and they must also watch the picture. This is why reading speed matters so much in timing. If a sentence has many characters but only a second or two to be shown, the viewer cannot keep up. The fix is usually to condense the wording, to split the text into more cues, or to extend the time where the speech and picture allow it. Rules of thumb for acceptable speeds differ by audience and by language, and are often set lower for children subtitles and for material aimed at less experienced readers.
Gaps, cuts and the picture
Timing is not only about the voice. The picture matters too. Many editors prefer that a cue does not run across a cut in the picture, because the eye tends to read a change in the image as a signal that the text may also have changed. Handling this well is part of why teams pay attention to shot changes.
Gaps between consecutive cues matter as well. A very small blank interval helps the viewer see that new text has replaced old text. Without it, two cues with similar wording can seem like one unchanged block.
Why subtitles go out of step
Even well-timed subtitles can fall out of step with the video. The usual causes are a different version of the video, a different frame rate, or an edit that shifted everything by a few seconds. Sometimes the offset is constant, and sometimes it grows as the video plays, which is called drift. The cause and the cure differ, and it helps to know which kind of problem you are looking at before you try to fix it.
Frame rate is a frequent culprit. A file timed for one rate and played against video at another may match at the start and gradually separate.
Timing for different situations
The same principles apply across many settings, with different emphasis:
- Recorded programs: Cues can be prepared in advance, checked and adjusted. This allows careful placement relative to speech and cuts.
- Live material: Text is produced as people speak, so it inevitably appears after the words. The unavoidable caption delay is a separate problem from timing a finished file.
- Short social clips: Videos watched with the sound off need text that keeps pace closely, since the viewer relies on it entirely.
- Cinema and large screens: Cue changes and gaps are read differently on a large screen, and cinema subtitles often follow stricter conventions.
Who sets the rules
Broadcasters, streaming services, film distributors and public bodies often publish their own guidelines on minimum and maximum durations, gaps and reading speeds. These differ, so a project should follow the specification given by whoever will use the subtitles. If no specification exists, a conservative approach is sensible: allow comfortable reading time, keep cues aligned with speech, and avoid very short cues.
Where the subtitles are meant to meet an accessibility requirement, remember that rules vary from place to place and change over time. A qualified professional or the relevant authority can advise on what applies to a particular project.
Checking timing
Timing is best judged by watching the video with the subtitles, not by looking at the numbers alone. A reviewer asks simple questions. Does the text arrive when the speech starts? Is every line readable at normal speed? Do cues change cleanly at cuts? Are there any places where the text stays or disappears at an odd moment? Tools can flag cues that are too short, too long or too fast, but they cannot tell whether the result feels natural. A person watching the finished video remains the most reliable test, and including timing in a regular quality checklist keeps the review consistent.