WCAG 1.2.6: Sign Language (Prerecorded)

Where 1.2.6 sits: the AAA layer of media accessibility

Guideline 1.2 is a ladder. At Level A, 1.2.2 Captions (Prerecorded) requires captions for prerecorded video. At Level AA, 1.2.5 requires audio description for blind viewers. Level AAA then adds three enhancements: sign language interpretation (this criterion), extended audio description (1.2.7), and a full text media alternative (1.2.8).

Because it is AAA, 1.2.6 is not part of the Level AA conformance most laws reference. W3C itself notes that AAA conformance is not recommended as a blanket policy for entire sites, because some content cannot satisfy every AAA criterion. Instead, teams typically apply 1.2.6 to high-value content: public announcements, emergency information, onboarding and training video, and anything aimed specifically at Deaf audiences.

Who this helps

What a conforming interpretation looks like

• →It covers all prerecorded audio in the synchronized media — dialogue from every speaker plus meaningful non-speech information such as who is speaking and important sounds.
• →It uses a sign language appropriate to the audience. ASL, BSL, Auslan, DGS, and others are distinct languages — an ASL interpretation does not serve a British audience.
• →It is performed by a qualified interpreter (or fluent signer) large and clear enough on screen to be read: hands, face, and upper body visible against a plain contrasting background.
• →It stays synchronized with the source audio, the same way captions must.
• →It can be 'open' (part of the video image, e.g. a corner inset) or 'closed' (a user-enabled synchronized stream or an alternate signed version of the video).

Implementation patterns

1. Offer a signed version of the video

The simplest conforming approach: produce a second render of the video with the interpreter composited into the image, and link it next to the original.

<video controls poster="/media/orientation-poster.jpg">
  <source src="/media/orientation.mp4" type="video/mp4" />
  <track kind="captions" src="/media/orientation.vtt"
         srclang="en" label="English" default />
</video>

<p>
  <a href="/media/orientation-asl.mp4">
    Watch this video with ASL interpretation
  </a>
</p>

2. A user-toggled, synchronized interpreter overlay

A “closed” interpretation keeps the main video clean for everyone else. Play a second, muted interpreter video in sync with the main one and let the user show or hide it.

<div class="player">
  <video id="main" controls src="/media/keynote.mp4"></video>
  <video id="signer" muted hidden
         src="/media/keynote-asl.mp4"
         aria-label="ASL interpretation"></video>
  <button id="toggle-sign" type="button" aria-pressed="false">
    Show sign language interpreter
  </button>
</div>

<script>
  const main = document.getElementById("main");
  const signer = document.getElementById("signer");
  const toggle = document.getElementById("toggle-sign");

  toggle.addEventListener("click", () => {
    const show = signer.hidden;
    signer.hidden = !show;
    toggle.setAttribute("aria-pressed", String(show));
    signer.currentTime = main.currentTime; // stay in sync
    if (show && !main.paused) signer.play();
  });

  // Mirror play, pause, and seeking
  main.addEventListener("play", () => !signer.hidden && signer.play());
  main.addEventListener("pause", () => signer.pause());
  main.addEventListener("seeked", () => {
    signer.currentTime = main.currentTime;
  });
</script>

3. Production notes for the interpreter video

The interpretation only works if it can be read. When compositing an inset, keep it large — W3C techniques suggest the signer occupy a meaningful share of the frame, not a thumbnail.

/* Corner inset that stays legible at typical sizes */
.signer-inset {
  position: absolute;
  right: 1rem;
  bottom: 3.5rem;            /* clear of the control bar */
  width: clamp(160px, 25%, 320px);
  aspect-ratio: 3 / 4;       /* head-and-torso framing */
  border: 2px solid #fff;
  border-radius: 0.5rem;
  background: #1e293b;       /* plain, contrasting backdrop */
}

How to test for 1.2.6

1. 1

   Inventory the synchronized media

   List every prerecorded video that has an audio track. Media that is a labelled alternative for text is exempt; everything else is in scope for AAA conformance.

2. 2

   Confirm an interpretation exists and is discoverable

   For each video, verify there is either an interpreter in the picture, a control that enables a synchronized interpretation, or an obvious link to a signed version adjacent to the video.

3. 3

   Check completeness

   Play the full video. The interpretation must run for all audio content — every speaker and meaningful sound — not just the intro or the main narrator.

4. 4

   Check legibility and synchronization

   At the sizes users actually watch (including mobile), the signer's hands and face must be clearly readable, and the interpretation must track the audio without drifting.

5. 5

   Verify the language matches the audience

   Confirm the sign language provided is the one used by the content's primary audience, and ideally have a fluent signer review quality — this is a human-judgment check no automated tool can make.

Common failures

• ✗Relying on captions alone and assuming they make sign language unnecessary — the two serve different needs and different criteria.
• ✗An interpreter inset rendered so small (or compressed so heavily) that handshapes, fingerspelling, and facial grammar are unreadable.
• ✗Interpretation that covers only part of the audio — for example, the host but not interview guests, or speech but not plot-critical sounds.
• ✗Providing the wrong sign language for the audience, such as ASL on content aimed at BSL users.
• ✗A signed version that exists but is buried — no link or control anywhere near the original video.
• ✗An interpreter overlay that drifts out of sync after seeking or pausing, so signs no longer match the audio.

Frequently asked questions