Matterhorn 13-007 · PDF/UA-1 UA1:7.3-6
Tables and text as pictures: when alt text is the wrong fix
A data table rendered as a picture, or body text set as an image, when real text and real table structure were available.
The condition, as the protocol states it. “A more accessible representation is not used.”
Matterhorn Protocol 1.1, checkpoint 13, index 13-007, section UA1:7.3-6. The protocol marks this as requiring human judgement.
What a screen reader actually does
A short alt text cannot substitute for a table of 200 numbers; the data is effectively lost.
What the condition is about
Some content arrives as a picture when it did not have to. The common cases:
- A table pasted as an image. A screenshot of a spreadsheet, or a table copied and pasted as a picture, sitting in a report whose every other table is real.
- Text set as an image. A quoted passage, a policy excerpt, a slide’s body text exported as a graphic, an email pasted as a screenshot.
- A list drawn as a diagram. Five steps in boxes joined by arrows, where the arrows only ever go one way.
PDF/UA-1 clause 7.3 asks that where a more accessible representation of content is available, it is used. A picture of a table is a figure; at best it gets a description. The same data as a tagged table gets rows, columns and headers that a reader can move through.
/Figure /Alt (Table of regional results for 2025 and 2026.) ← what it is
/Table ← what it should be
├── /TR TH "Region" TH "2025" TH "2026"
├── /TR TH "North" TD "120" TD "135"
└── …What breaks
A screen reader user meets the figure, hears its description, and moves on. If the description is short, the data is gone. If it is long, it is read as one unbroken stream that cannot be navigated — a listener who wants the 2026 figure for the South region has to hear everything before it, and again if they missed it.
It is not only screen readers. Text in a picture does not reflow or enlarge cleanly, cannot be searched, copied, translated or read by text-to-speech, and does not change colour in high-contrast modes. Everything WCAG 1.1.1 asks a text alternative to make possible is easier with the text itself.
Where it comes from
Paste as picture. A table copied from a spreadsheet into a word processor or a slide as an image, often to stop it reformatting.
Screenshots of other documents. The quickest way to include something from another system is to capture it.
Design. Text set inside a graphic for typographic control, or a slide whose body is a single image.
Merging. A born-digital report with a scanned appendix or a pasted-in page from elsewhere. A scanned page is a job of its own — see Remediate a scanned PDF.
How to check it free
Search for something you can see. Ctrl/Cmd + F for a word or number visible in the table. No match on that page means it is not text.
Try to select it. If the text-selection cursor selects the whole block as one object, or nothing, it is a picture.
Acrobat Pro’s Reading Order tool. Select Show Tables And Figures: a “table” outlined as a figure is this condition.
Acrobat’s Accessibility Checker will not find it. Its Image-only PDF rule, as Adobe documents it, is about a document that “appears to contain text, but doesn’t contain fonts” — a whole file of images, not one image in a document full of text. Its Figures alternate text rule passes as soon as the picture has a description. PAC and veraPDF check figures for a description too, and are satisfied the same way.
How to fix it
In the source. Replace the picture with the real thing: paste the table as a table, type the passage as text, turn the flowchart into a numbered list, then re-export. Nothing done to the PDF afterwards is as good, and nothing else survives the next revision. See Excel to accessible PDF, Word to accessible PDF and Tag tables correctly.
In Acrobat Pro, honestly: there is no conversion. Acrobat has no command that turns a picture of a table into a tagged table. Its text recognition (All tools → Scan & OCR → Recognize Text → In this file) is what Adobe’s Create and verify PDF accessibility page points to for scanned images; even where it recovers the words, the table structure still has to be built by hand with the Reading Order tool’s Table button, and every cell checked, because recognition errors look exactly like data.
When the source is gone and the data is not. If the figures exist somewhere — a spreadsheet, a published dataset — build a real table from them in any authoring tool, export it as a tagged PDF and insert it as an appendix page, with the figure’s description saying where the table is. Adobe’s Creating accessible PDFs page notes that Acrobat adds an inserted page’s tags to the end of the tag tree, which for an appendix is where they belong. It is a patch, and a re-export is still the fix.
What Taggart does
Taggart has no check for this condition, so it produces no finding and does not appear in the review queue. It is on the list of conditions Taggart asks you to confirm before it writes the PDF/UA identifier, under one box for all of them — I have checked these conditions by hand, or with another validator — and the identifier waits until it is ticked.
Where it can mislead is one step earlier. A picture of a table with no description is reported as 13-004, and the review queue gives it a description box like any other figure. The box lists what the document already says about the image — its caption, the line below it, text painted inside its area. A pasted table image has nothing painted inside its area to offer, because its numbers are pixels; if there is nothing around it either, the queue says so rather than inventing a description. Writing a sentence into that box will pass 13-004 and leave this condition exactly where it was.
So for a table or a passage that is a picture, the right answer in the queue is not a description. Mark it Needs work, say why in the justification — table pasted as an image; source table to be re-exported — and fix it at source. The decision and its reason go into the remediation report.
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.
Whatever Taggart changes, the page still looks identical. Fixes edit the PDF object model, not the content streams, and every remediation is re-rendered and compared against the original at structural similarity ≥ 0.999 before you get the file back. A tool that quietly reflows your document is worse than no tool.
Frequently asked questions
Can a long alt text stand in for a table?
Not one of any size. Alt text is read as a single block: no rows, no columns, no headers announced as you move, no way to go back to the third figure in the second column. A table of five numbers can survive being turned into a sentence; a table of fifty cannot.
What counts as a more accessible representation?
Whatever the content already is, underneath the picture. Data in rows and columns is a tagged table. Prose is text. A sequence of steps drawn as a flowchart may well be a list. The test is whether the information could have been set as text and structure, and was set as an image instead.
Are photographs and charts this condition?
No. A photograph has no text representation, and a chart's shape is the point of it — those are figures with descriptions. A chart whose underlying numbers matter can carry the numbers in a real table alongside it, which is a better answer than a long description.
Is a scanned page this condition?
A scan is the extreme case of it, but it has its own route: text recognition first, then tagging. See the scanned-PDF guide. This page is about images placed in an otherwise digital document.
Where this sits in the standards
| Standard | Reference |
|---|---|
| Matterhorn Protocol 1.1 | Checkpoint 13 (Graphics), index 13-007 |
| PDF/UA-1 (ISO 14289-1) | Clause 7.3-6 |
| WCAG 2.1 | 1.1.1 Non-text Content — Level A |
| EN 301 549 / Section 508 / ADA Title II | All three point at WCAG 2.1 Level AA for non-web documents, so a Level A or AA criterion here is in scope for each of them. |
Related conditions
- 13-004 Figure tag alternative or replacement text missingA Figure element has neither /Alt nor /ActualText. Every Figure must have one or the other; if the image is decorative it should be an artifact instead of a Figure.
- 13-006 A graphic with meaning only as part of a group is tagged on its ownA chart built from dozens of separate vector paths has been tagged as dozens of Figures instead of one. Each fragment is meaningless alone.
- 15-002 A column has a header cell, but that header cell is not tagged as a headerThe top row labels the columns but is tagged as ordinary data cells.
- 13-008 ActualText not present when a Figure is intended to be consumed primarily as textA drop cap, a logo that is a word, an image of a signature block — content that is text — needs ActualText so it flows into the text stream.
- 31-027 A font dictionary does not contain the ToUnicode entry and no exemption appliesA ToUnicode CMap is required unless the font uses MacRoman/MacExpert/WinAnsi encoding, or is a Type 1/Type 3 whose glyph names are all in the Adobe Glyph List or the Symbol set, or is a Type 0 whose descendant uses a standard Adobe character collection, or is a non-symbolic TrueType.
Other conditions in checkpoint 13
Checkpoint 13 covers graphics. These are its other failure conditions; the ones without a link are in the rule set but do not have a written page yet.
- 13-001 Graphics objects are not tagged with a Figure tagAn image or vector drawing that carries meaning is not inside a Figure element. Either it is untagged, or it has been swept into a paragraph.
- 13-002 A link with a meaningful background lacks alternative text for bothAn image link — a button, a logo that navigates — needs alternative text that conveys both what the graphic shows and where the link goes.
- 13-003 A caption is not tagged with a Caption tag"Figure 3: Quarterly revenue" is tagged as a paragraph rather than as a Caption associated with its figure or table.
- 13-004 Figure tag alternative or replacement text missingA Figure element has neither /Alt nor /ActualText. Every Figure must have one or the other; if the image is decorative it should be an artifact instead of a Figure.
- 13-005 ActualText used for a Figure for which alternative text is more appropriateActualText replaces the content for text-extraction purposes; Alt describes it. A photograph needs Alt. A drop cap or a glyph-as-image needs ActualText.
- 13-006 A graphic with meaning only as part of a group is tagged on its ownA chart built from dozens of separate vector paths has been tagged as dozens of Figures instead of one. Each fragment is meaningless alone.
- 13-008 ActualText not present when a Figure is intended to be consumed primarily as textA drop cap, a logo that is a word, an image of a signature block — content that is text — needs ActualText so it flows into the text stream.
Check your own file. Taggart’s validator is free and unlimited — 75 of the 87 Matterhorn conditions software can decide, with the other 12 listed for you to confirm; no watermark, no expiry, and no account needed up to 60 pages.
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 . The Acrobat steps follow Adobe’s documentation and were not walked through by hand. Last updated .