Captions and audio description are production work, not a final checkbox

Both approaches to accessibility features work. What differs is what they cost you, and the cost is what this sets out.

Captions and subtitles are not the same thing

Subtitles assume the viewer can hear and renders speech in another language, while captions assume the viewer cannot hear and must convey everything meaningful on the soundtrack. That means captions include speaker identification, significant sound effects, music cues and tonal information that a subtitle would leave to the audio. Using a subtitle track as a caption track is a common shortcut that leaves deaf and hard-of-hearing viewers without information the scene depends on.

Describing sound well requires judgement, since captioning every noise is unreadable and captioning too little removes the atmosphere entirely. Good caption writing is therefore a craft with its own conventions rather than a transcription exercise.

Timing and reading speed decide comprehension

A caption must be on screen long enough to read and must not run so far ahead that it reveals a line before it is spoken. Reading speed limits force condensation, so captions are frequently edited rather than verbatim, and that editing is an interpretive act. Placement matters as well, since a caption covering a face or an on-screen sign creates a new problem while solving another.

Rapid overlapping dialogue is the hardest case and often cannot be fully represented, which is a genuine limitation rather than laziness. Automatic caption generation has improved substantially for clear speech and remains unreliable for accents, overlapping voices and unusual terminology.

Audio description is written into gaps

A description track narrates visual information for blind and partially sighted viewers, and it can only speak where the existing soundtrack leaves room. That constraint shapes everything, since a densely scored action sequence may allow only a few words while a quiet scene allows a sentence.

The writer must choose what matters most, which means deciding whether a facial expression, a location detail or an action is the essential information. Description is then performed, and the choice of voice, pace and level determines whether it sits inside the film or on top of it. Extended description that pauses the programme exists for some material and is not usable for live or broadcast content.

Planning it early is cheaper and better

When access work happens after delivery, the team has no contact with the production and must infer intent from the finished file alone. Done during post, description and captioning can draw on the script, the sound mix and the people who made the decisions being described. Mixers can also leave deliberate space in the soundtrack if they know description is coming, which improves the result at no real cost.

The same applies to on-screen text, where designing titles with caption placement in mind avoids collisions later.

Productions that treat access as a delivery requirement rather than a craft consistently produce worse results for the same money.

Requirements differ and vary widely

Obligations to provide captions and description are set nationally and differ substantially in scope, thresholds and enforcement. Some regimes apply to broadcasters but reach streaming services only partially, which is why coverage on a single service varies by territory. Coverage also varies by title, since older licensed material may have no description track and creating one is a fresh cost.

Anyone relying on these features should check availability per title rather than assuming a service provides them universally. Advocacy organisations in several countries publish comparative information, which is generally more current than any summary written elsewhere.

The features are used far beyond their intended audience

Caption use is now widespread among viewers with no hearing loss, driven by dense mixes, quiet dialogue and watching in noisy or shared environments. That has changed the argument, since a feature used by a large proportion of an audience is no longer easy to characterise as an optional extra.

Some productions have begun considering caption quality as part of presentation rather than compliance, which is a meaningful shift. Audio description has a smaller general audience but is valued by viewers doing other things and by those who find it clarifies dense visual storytelling. Designing for access reliably improves the product for everyone, which is an old lesson that each medium seems to learn separately.

Side by side

The takeaway

Access work done during post is craft. Done after delivery it is only compliance.

Watch the transitions. That is where the argument of a film usually is.