PDF/UA-1 · CLAUSE 7.21.6
Font is not embedded or is inconsistently encoded
A font used for rendering is not embedded, is missing glyphs, or has encoding tables that do not let a reader map character codes to glyphs reliably.
Why it matters
Text may render with substituted glyphs, and text extraction can produce wrong characters, so both sighted readers and screen reader users can get corrupted text.
How to fix it
- Embed all fonts, including subsets, when exporting the PDF.
- Avoid fonts with broken or missing cmap tables; substitute a well-formed font in the source document.
- Check the export preset: 'optimised for size' presets often drop font embedding.
What the standard says
All non-symbolic TrueType fonts shall have either MacRomanEncoding or WinAnsiEncoding as the value for the Encoding key in the Font dictionary or as the value for the BaseEncoding key in the dictionary which is the value of the Encoding key in the Font dictionary. In addition, no non-symbolic TrueType font shall define a Differences array unless all of the glyph names in the Differences array are listed in the Adobe Glyph List and the embedded font program contains at least the Microsoft Unicode (3,1 – Platform ID=3, Encoding ID=1) encoding in the 'cmap' table
Check your own PDF
Validate a file against PDF/UA-1 and see whether this rule fails in your document.
curl -sS -X POST "https://api.ua1.dev/api/validate?format=compact&profile=ua1" \ -F "file=@document.pdf"