Reference XObjects are not allowed

PDF/UA check REFERENCE_XOBJECT · Structure & other · severity High · PDF/UA ISO 32000-2 deprecation

↩ Back to the rules index · WCAG cross-reference

Why it matters#

Reference XObjects pull content from external PDFs, which breaks accessibility. The referenced content's tags and reading order are not part of the host document.

What you will see#

The document pulls content from an external PDF via a Reference XObject, deprecated in PDF 2.0.

How to fix#

Remove the Ref entry from each Reference XObject, converting it to a regular Form XObject. Embed the previously referenced content directly in the document and tag it in the structure tree.

Step-by-step guide#

1. Locate the offending Form XObjects: each one carries a /Ref entry pointing at content in an external PDF (the EQChecker report gives the count; Acrobat Preflight > options menu > Browse internal PDF structure lets you find them under page Resources > XObject).
2. Because the proxy content inside the XObject stream is what viewers already render, the standard fix is to delete the /Ref entry, converting each Reference XObject into an ordinary Form XObject. Acrobat has no menu or Preflight fixup for this - the file was generated with deprecated Reference XObjects, and the reliable fix is regenerating it from source with the artwork embedded directly.
3. If the proxy content is a low-fidelity placeholder rather than the real artwork, instead embed the referenced page content directly into the document and repaint it in place of the Reference XObject.
4. Tag the resulting content in the structure tree if it is real content, or mark it as an artifact if it is decorative - previously the referenced content carried no tags in the host document at all.

Verify: Re-run the EQChecker scan and confirm REFERENCE_XOBJECT passes with a zero count. A Preflight custom check for Reference XObjects (or a PDF/A validation, which also prohibits them) should find none remaining.

Note: Reference XObjects are deprecated in ISO 32000-2 (PDF 2.0). They import appearance from an external file, so the imported content's tags and reading order can never be part of the host document's structure tree - which is why they are incompatible with accessible PDF regardless of how good the external file is.

WCAG success criteria#

No WCAG success criterion maps cleanly to this check — it is a PDF/UA structural requirement. The PDF/UA clause above is the normative reference.

Standards#

This check contributes to the following standards:

PDF/UA