BREYDON’s methods of captioning internet

First published April 2020.
Updated 03 March 2022.

My approaches to captioning on BREYDON’s reflect particular accessibility priorities, while being heavily influenced by my use of Org for writing.

1. Placement

To ensure no‐one misses out on them, image descriptions and credits stay out of the (often heavily obscured) metadata and instead feature in the main flow of t’ text. In most media the descriptions are formatted as quotes — contributors to the conversation.

Where practicable, the picture itself follows as a captioned figure. The figure caption can serve as an index listing.

It took some fiddling to find a compromise that would suffice across many media, while keeping the single, shared source file succinct. This is where I’ve settled for now:

[​[filepath][A one‐line description of the image.]]
Role: Person. Thing copyright who, used under what licence.

Any subsequent detail for the image description.

#+name: fig:unique_identifier
#+caption: Event or subject. Location, month year.

This same approach can be applied to music notation, rubrics, and other means of visual representation, regardless of whether an actual image file is involved or not.

2. Dynamic subtitles

We can go beyond the wall of text, of course.

Synchronised captions for speech and other sounds are possible if the audio is presented like a film clip.

Once the audio is mastered, a written copy of the text goes into a subtitle file, where I gradually break the piece down into time‐stamped chunks. For a basic monologue, the simplistic SubRip format is adequate. Working in the absence of dedicated subtitling software, Web Video Text Tracks (WebVTT) are even easier to edit.

Next, the sounds and the subtitles are packaged together. Some might prefer if helpful imagery chaperones the subtitles, despite video files incurring bigger demands. But to conserve space and energy, it should be possible to distribute all the subtitle tracks that anyone could want in a humble Matroska audio container.

Alternatively, app‐aspirant H.T.M.L. tags can combine the audio with WebVTT on a webpage, like so:

<figure id="videoContainer" data-fullscreen="false">
  <video id="video" controls preload="metadata" poster="whatever.jpg">
    <source src="whatever.ogg" type="video/ogg">
      Even if the audio will not play on this page, you can still
      <a href="whatever.mka">download a captioned recording</a>.
    <track label="English" kind="captions" srclang="en" src="whatever-en.vtt" default>
  <figcaption>A close‐captioned recording of whatever.</figcaption>

3. Multilingualism

For publishers with adequate resources, captioned readings could easily extend to subtitles in translation and footage of signing interpreters!