Thoughts on subtitles, part 2

Subtitling is seriously undervalued in the South African film and television industry. This is the second in a series of blog posts covering subtitling concerns.

Timing and synchronisation

A fundamental function of subtitling is to reduce frustration caused to hearing-impaired viewers by being faced with silent moving mouths. Therefore, all obvious speech should have some form of subtitle accompaniment.

Eye movement research shows that hearing-impaired viewers make use of visual cues from the faces of television speakers in order to direct their gaze to the subtitle area. Therefore, subtitle appearance should coincide with speech onset and subtitle disappearance should coincide roughly with the end of the corresponding speech segment, since subtitles remaining too long on the screen are likely to be re‑read by the viewer.

The same rules of synchronisation should apply with off-camera speakers and even with off‑screen narrators, since viewers with a certain amount of residual hearing make use of auditory cues to direct their attention to the subtitle area.

Leading and lagging

The target point for synchronisation should be at naturally occurring pauses in speech-sentence boundaries, or changes of scene. However, there are bound to be cases where this is either impractical or inapplicable.  

Shot changes

Besides the general recommendation for subtitle/speech synchronisation, there are certain other aspects of the television picture which influence subtitle timing. Subtitles that are allowed to over‑run shot changes can cause considerable perceptual confusion and should be avoided. Eye-movement research shows that camera‑cuts in the middle of a subtitle presentation cause the viewer to return to the beginning of a partially read subtitle and to start re‑reading.  In practice, it is recognised that the frequency and speed of shot changes in many programmes present serious problems for the subtitler. A subtitle should, therefore, be “anchored” over a shot change by at least one second to allow the reader time to adjust to the new picture. Shot changes normally reflect the beginning or end of speech. The subtitler should, therefore, attempt to insert a subtitle on a shot change when this is in synchrony with the speaker.

Dealing with visual edits

  1. Avoid inserting a subtitle less than one second before a visual edit and removing a subtitle less than one second after an edit.
  2. Attempt to insert a subtitle in exact synchrony with an edit.
  3. A decision to segment a single sentence into more than one subtitle, to be placed around a camera‑cut, should depend on whether the sentence can be segmented naturally and on whether the resulting subtitles can be allowed sufficient display time.

Tone of voice

Where tone of voice is particularly critical to meaning, and facial expression and body language are inadequate to convey the tone, the use of “(!)” and “(?)” immediately following speech can indicate sarcasm and irony as shown below:

No, no. You’re not late (!)

Rapid dialogue techniques

All obvious speech should be accompanied by subtitle information, but under conditions of rapid dialogue, several short subtitles displayed in rapid sequence can result in staccato or ‘machine-gun’ effect. There are two possible solutions for this,  double-text and add-ons.


  1. Double-text can be used when two characters or more speak simultaneously.
  2. Add-ons should normally be preferred when two or more characters speak consecutively and time does not allow individual subtitles.

The total length of either double-text or add-on sequence should never exceed four lines.

Double text

Double text is normally used when more than two characters speak simultaneously and contradict one another.

Double-text can be an infringement of the synchronisation recommendation: the appearance of each subtitle should coincide with the beginning of the corresponding speech segment. 

Have you had lunch?
Yes.                  No.

Both people’s speech is contained within one subtitle.

Add-ons or cumulative titles

The second solution is to use “dynamic” text known as “add-ons” or “cumulative titles”. This is most effective when the two subtitles fit naturally together, for example in a question and answer sequence, or providing the punch line of a joke. In this technique the second part of the title is added on to the first part.  Thus, the appearance of the second part, or “add-on” of the title can coincide with the onset of the second utterance, while the subtitle corresponding to the first utterance remains on screen. In this way, the staccato effect is diminished while still preserving a natural relationship between speech onset and subtitle presentation.

For example, building a full subtitle in three steps:

No... No..
No... No... But that isn’t what I asked for.

A further advantage of add-ons is that they appear more natural when the two corresponding speakers are not shown in the same camera shot.

However, additional reading time is gained only if the gaze of the viewer remains on the subtitle area throughout its presentation. If the gaze of the viewer returns to other parts of the picture before the add-on appears, as can occur among unusually fast readers, the subsequent add-on can cause some perceptual confusion.

In both double‑text and add-ons, the second part of the text should normally appear on the line immediately beneath the first part.

Sourced from the ITC Guidance on Standards for Subtitling - February 1999