Matterhorn 05-001 · PDF/UA-1 UA1:7.1-7

Video or audio in a PDF with no transcript

Embedded media carries information in audio with no transcript or equivalent text in the document.

What a screen reader actually does

A deaf or deafblind reader gets nothing from the media.

What a media annotation is

PDF plays audio and video through annotations that reserve a rectangle on the page and point at the media. Three kinds exist: Screen annotations driven by a rendition action, which are the current mechanism; Movie annotations, deprecated in PDF 2.0; and RichMedia annotations, an Adobe extension standardised in PDF 2.0. Acrobat Pro adds video and sound through All tools → Add rich media, from a file or from a URL (Adobe: Adding multimedia to PDFs).

A Screen annotation carries two descriptions, and neither is what this condition asks for:

40 0 obj
<< /Type /Annot /Subtype /Screen
   /Contents (Induction video)                      % what the annotation is
   /A << /S /Rendition /AN 40 0 R /OP 0
         /R << /S /MR
               /C << /Type /MediaClip /S /MCD
                     /CT (video/mp4)
                     /Alt [ () (Site induction, 2 minutes) ]   % what the clip is
                     /D 41 0 R >> >> >>
>>

/Contents is the annotation’s description, required by 28-004. /Alt on the media clip is the clip’s description, required by 28-015. Both say what the media is. Condition 05-001 is about what is in it: if the clip’s audio carries information, that information has to exist somewhere in another form — normally text.

What the reader gets

A screen-reader user reaches the annotation and hears its description; if the media plays, they hear the soundtrack, so the audio itself is not their problem — though anything shown only on screen is, which is WCAG’s audio description requirement rather than this condition.

A deaf reader sees the video and hears nothing. Without captions or a transcript, every spoken instruction in it is lost. A deafblind reader on a braille display gets the description — Induction video — and that is the whole of the content as far as the document is concerned.

The standards table below maps the condition to WCAG 1.2.1, for audio-only content. Video with sound also falls under 1.2.2, captions, and the audio description criteria; Adobe’s Accessibility Checker cites the same group of criteria for its Tagged multimedia rule.

Where it comes from

Media often gets added late. A training document is written and tagged properly, then someone adds the induction video in Acrobat the week before release. The tagging was checked; the video arrived afterwards with a one-line description and no transcript, because nobody’s checklist had a step for it.

It also comes from the video team rather than the document team: the clip exists, the captions exist as a sidecar file for the website, and the PDF embeds the clip without them. Adobe’s documentation for adding video to a PDF describes poster images, playback controls, trimming and chapter points, and no caption setting, so do not count on captions that are not burned into the picture.

How to check it free

There is no machine test: whether audio carries information, and whether the document already says the same thing in text, are both judgements.

Find the media. This lists every media annotation with its page and its description, using Python and the free pikepdf library:

import sys, pikepdf

MEDIA = {"/Sound", "/Screen", "/Movie", "/RichMedia"}
pdf = pikepdf.open(sys.argv[1])
for n, page in enumerate(pdf.pages, 1):
    for a in page.get("/Annots", []):
        if str(a.get("/Subtype")) in MEDIA:
            print(f"page {n}: {a.Subtype}  Contents={a.get('/Contents')}")

In Acrobat Pro, the Accessibility Checker’s Tagged multimedia rule covers the same objects from the tagging side (All tools → Prepare for accessibility → Check for accessibility; Adobe: Create and verify PDF accessibility).

Play each one with the sound on and the text open. For every spoken sentence, sound effect or audible cue that matters, find where the document says the same thing. If the answer is “nowhere”, the condition fails. If the answer is “in the paragraph above, mostly”, decide whether mostly is enough — a summary that omits the safety instructions is not a transcript of a safety video.

How to fix it free

Write the transcript. Many video platforms and editors produce a caption file automatically; correct it and you have a transcript. Include speaker changes and meaningful non-speech sound — [alarm] — and, for video, enough description of what is shown to follow it.

Put it in the document. The safest place is ordinary text right after the media in the reading order, under a heading such as Transcript, so every viewer and every reader gets it. Add it in the source document and export again; tag it like the rest of the text. If space is the problem, put it in an appendix and say so in the media’s description: Induction video, 2 minutes. Transcript in appendix B.

Add captions to the video itself if it is to be watched in the PDF. Burn them into the picture before inserting it, since Adobe documents no caption option for inserted video, then replace the clip: select it with the Select object tool and press Delete, and insert the captioned version from All tools → Add rich media → Video.

What Taggart does

Taggart has no check for this condition. It does not play media, transcribe audio or compare a soundtrack with the text, so the condition never appears in the findings or the review queue.

What it does check about the same object is the plumbing. A Screen, Movie or RichMedia annotation that is not inside an <Annot> element is reported as 28-002, and the annotation-structure repair — a safe one, applied by Fix All Safe — puts it there, placed by where it sits on the page: inside the element whose text it covers, or just before the first element below it, and at the end of the tree only when neither exists. One with no /Contents and no alternate text on its element is 28-004; Taggart does not write that description itself. And it follows the rendition action to the media clip: a clip with no /CT is 28-014, and one with no /Alt is 28-015. Neither has a repair in Taggart.

None of that is a transcript. On a test file where a properly tagged and described video sat on a page with no transcript anywhere — alongside a handful of other problems of the same judgement kind — Taggart reported no findings. What it did not do was claim conformance: its statement stopped at machine checks passed, with 05-001 among the conditions Taggart does not check that nobody had confirmed, and asked to declare conformance it left the PDF/UA identifier out. It writes the identifier once a person confirms that list — the checkbox on the application’s conformance step. The check above is what that confirmation should mean for any document with media in it.

How Taggart handles it

No software can decide this one, and Taggart has no check for it, so it never appears among the findings or in the review queue. It is on the list of conditions Taggart asks you to confirm by hand before it writes the PDF/UA claim into the file — the claim waits until you have.

Where this sits in the standards

Related conditions

• 05-002 Audio annotation present, but content not available in another formA Sound annotation carries information with no transcript.
• 03-002 Multimedia objects contain flickering contentEmbedded video or animation flashes more than three times per second.
• 28-004 An annotation has neither a Contents entry nor an alternative descriptionEvery non-widget annotation needs an accessible description, either in its own /Contents or in the /Alt of the structure element that encloses it.
• 28-015 Alt entry is missing from the media clip data dictionaryMedia clips require an /Alt array giving a text description.

Other conditions in checkpoint 05

Checkpoint 05 covers sound. These are its other failure conditions; the ones without a link are in the rule set but do not have a written page yet.

• 05-002 Audio annotation present, but content not available in another formA Sound annotation carries information with no transcript.
• 05-003 JavaScript uses beep but provides no other notificationForm validation signals an error only with app.beep().

This condition needs a person’s judgement, so there was nothing for a tool to check. Written from ISO 14289-1 and the Matterhorn Protocol on 2026-10-02. The Acrobat steps follow Adobe’s documentation and were not walked through by hand. Last updated 2026-10-10.