Whether it’s for videos on the internet or cinematics in video games, captions are an essential accessibility feature. Note that we are talking about “closed captions” here, which are not about translating content—these are subtitles.

For hard-of-hearing and deaf people of course, but also for people for whom processing audio might not be possible (such as those without headphones in a loud environment) or overwhelming (which can be the case for people on the autistic spectrum). They are also very handy for non-native speakers for whom understanding content might be easier when seeing it written rather than just spoken out.

It turns out that authoring good captions is actually surprisingly difficult, and the quality from source to source greatly varies. Here is a collection of tips to make captions as useful as possible:

  • Captions should usually live in the safe area of a 16:9 screen resolution, at the bottom of the screen. They might be temporarily moved when obscuring relevant visual content such as embedded text.

  • Captions are meant to be read, and therefore their size matter. They should be big enough to be readable at most distances but not too big that they would need to be refreshed too often.

  • Like for any text content, contrast is key. The ideal colors for captions on a solid dark background are white, yellow and cyan. Colors can also be used to denote different speakers within a conversation, which can really help understanding.

  • The length of captions should be kept short (~40 characters) and the text should not stick to the sides since differences in screen calibrating could cut the edges off. A caption should usually not exceed 2, maybe 3 lines.

  • Captions should be displayed for 1 or 2 seconds and changes of captions should come with a brief (200—300ms) uncaptioned pause to make sure the reader can acknowledge a change of text even when lines look alike (length, etc.).

  • Language-specific typographic rules should be respected. Words should be broken where possible according to the language they are depicted in, and sentences should be split on punctuation as much as possible.

  • Special care can be taken to make sure not to spoil upcoming events before they appear on screen. Nothing like knowing what happens before it actually does because the caption was too revealing.

  • Important sound effects and subtility (such as tone, emotions, loudness, music…) should be explicitly mentioned. Same thing if the sound/dialogue comes from something off-screen.

As you can see, there are a lot of things to consider to make captions accessible. Some content might be easier to caption that others (single speaker, few editorial cuts, no sound effects or music…). The more attention is devoted to captions, the more accessible the content becomes. It is particularly critical when the main content of a given page or product is provided through videos (movie, series, screencasts…).