Add a sign language video stream in SMIL 1.0
SM13 is a sufficient technique when used with G81, and what it delivers is sign language interpretation rather than captions. That distinction is the whole point of the technique and it is easy to lose, because the two accommodations serve overlapping audiences and are not interchangeable. The method is two regions playing in sync, one holding the content video and one holding a video of an interpreter. G81, the partner technique, sets the terms for the interpreter video, and its own title carries the requirement. The interpreter is displayed in a different viewport or overlaid on the image by the player. Both of those leave the viewer some control. An interpreter burned permanently into the video frame is the one arrangement G81 does not describe, because a viewer who does not need it cannot move it or turn it off. Today the same thing is a second video track, or a signed version published alongside the standard one.
How we find it in an audit
This is entirely manual, and the first question is whether a signed version exists at all, which for most sites answers itself. Where one does exist we check three things. Does it stay in sync with the content, is the interpreter large enough and lit well enough to read, and can a viewer find the control that brings it up. Then we ask a Deaf reviewer, because whether an interpretation is good is not a question a hearing auditor can answer from a checklist. No scanner has any part in this.
How affected users experience it
For many Deaf people a national sign language is their first language, and written text is their second. That changes what captions are. Reading captions in a second language, at the speed a film runs, while also watching the film, is work. An interpreter carries tone, emphasis and register that caption text flattens into plain sentences. That is why sign language interpretation is its own accommodation with its own technique rather than a nicer version of captions, and why offering captions does not answer the need this technique exists to meet.
Passes vs. fails
Passes
<region id="main"/>
<region id="signer"/>
<par>
<video src="briefing.mp4" region="main"/>
<video src="briefing-signed.mp4" region="signer"/>
</par>
<!-- Two synchronized streams, and the player decides where the interpreter sits. -->Fails
<region id="main"/>
<par>
<video src="briefing.mp4" region="main"/>
</par>
<!-- Captions may exist. No interpretation does. -->Other ways to satisfy this rule
4 guides on this site are filed under 1.2.6 Sign Language (Prerecorded). W3C lists this one as sufficient for that rule only alongside G81, so the pair is what passes and neither half does on its own.
This guide is our interpretation of W3C technique SM13: Providing sign language interpretation through synchronized video streams in SMIL 1.0. W3C publishes its techniques as guidance rather than as the standard, and says so on every one of them. The success criterion is what conformance is measured against, and a technique is one documented way to meet it.