Captioning

All time-based media (i.e. video and audio) in ALEX requires closed captioning. This is not just an inclusive practice generally but an important part of our accessibility requirements.

Writing captions

When creating caption files, make sure they include:

  • Accurate spelling
  • Correct grammar and punctuation
  • Important non-speech sounds
  • Verbatim text (as close as possible, within reason)

If you’re unsure about how to format certain phrases (e.g. “health care” vs. “healthcare”), consult the project’s script as well as the ADS guidelines on formatting and vocabulary.

Formatting captions

When formatting captions:

  • Don’t exceed about 42 characters per line, including spaces
  • Don’t exceed two lines per caption

Break lines and captions at natural linguistic boundaries that follow the rhythm of speech. When breaking a caption into two lines, aim for lines of comparable length. Avoid splitting phrases, names, or compound ideas across lines or between captions (for example, separating articles from the words they modify).

Non-speech sounds

Time-based media in ALEX often uses sound effects and music to enhance the experience. Generally, this audio is not essential to comprehension. However, when it is helpful to caption non-speech sounds, include brief descriptors in the caption file wrapped in square brackets.

Examples

[wind]
[cat meows]
[phone ringing]
[calm music]

Timing captions

Captions should be tightly synchronized to the audio while remaining on screen long enough to read:

  • Captions should be timed against a frame rate matching the source media’s frames-per-second, or 30 frames-per-second for audio-only content
  • Target a reading speed of 17-20 characters per second for each caption
    • 1 line (~42 characters): ~2-2.5 seconds
    • 2 lines (~84 characters): ~4-5 seconds
  • The minimum on-screen duration for a caption is 1 second
    • Short audio files (e.g., “Okay!”) may need to be lengthened during editing to meet minimum duration requirements
  • Gaps between captions should be 0 or ≥ 400 ms (12 frames at 30fps) to avoid flickering