Checkpoint 30: XObjects

PDF documents must properly handle XObjects (external objects) for accessibility. Reference XObjects are prohibited, and Form XObjects containing marked content identifiers (MCIDs) must be uniquely referenced to maintain proper document structure.

What This Means

XObjects are reusable content objects in PDFs. The two types relevant to this checkpoint are:

Reference XObjects

Reference XObjects allow one PDF to reference content from another PDF file. While this can reduce file size and enable document modularity, it creates accessibility problems:

The referenced content exists outside the main document
Assistive technology may not be able to access the external content
The document structure becomes dependent on external files
If the referenced file is unavailable, content is lost

PDF/UA prohibits Reference XObjects entirely.

Form XObjects

Form XObjects are reusable content containers within a PDF. They're commonly used for:

Repeated graphics (logos, watermarks)
Header/footer content
Complex illustrations used multiple times
Pre-rendered content blocks

Form XObjects can contain Marked Content Identifiers (MCIDs), which link content to the document's tag structure. Each MCID must uniquely identify a piece of content.

The problem: If a Form XObject with MCIDs is used more than once, those MCIDs would appear multiple times in the document. This breaks the assumption that each MCID uniquely identifies content, causing:

Confusion in the tag structure
Incorrect reading order
Assistive technology errors
Broken content-to-tag mapping

Why It Matters

These XObject restrictions protect document structure integrity:

Reference XObjects:

External dependencies make documents fragile
Users may not have access to referenced files
Screen readers cannot reliably process external content
Document portability is compromised
Validation becomes impossible without external files

Form XObjects with MCIDs:

MCIDs must be unique for proper tag mapping
Duplicate MCIDs create ambiguous content references
The tag structure becomes inconsistent
Screen readers may announce content incorrectly or incompletely
Navigation through the document becomes unreliable

Consider a logo Form XObject with an MCID linked to a Figure tag with alt text. If this Form XObject appears on 50 pages, those 50 instances would share one MCID. Assistive technology expects each MCID to appear once, creating confusion about which instance is being referenced.

Common Violations

The Matterhorn Protocol defines two failure conditions for XObjects. Both are machine testable.

30-001: Reference XObject Is Present (Machine Testable)

What's Wrong: The PDF contains a Reference XObject, which references content from an external PDF file. This is prohibited in PDF/UA.

How to Identify:

PDF/UA validators automatically detect Reference XObjects
The violation appears in validation reports as a structural error
Reference XObjects are relatively rare in modern PDFs

Technical Context: A Reference XObject contains:

<< /Type /XObject
   /Subtype /Form
   /Ref << /F (external-file.pdf) ... >>
>>

The presence of the /Ref key with a file reference indicates a Reference XObject.

Why This Happens:

PDF creation tools that optimize for file size
Document assembly workflows that link rather than embed
CAD or technical drawing exports with external references
Legacy PDF workflows from when Reference XObjects were common

30-002: Form XObject Contains MCIDs and Referenced More Than Once (Machine Testable)

What's Wrong: A Form XObject that contains Marked Content Identifiers (MCIDs) is referenced multiple times in the document. This creates duplicate MCIDs, breaking the tag structure.

How to Identify:

PDF/UA validators detect this automatically
The validator will identify the specific Form XObject
Reports typically indicate which pages or locations reference the object

Technical Context: A Form XObject with MCIDs looks like:

<< /Type /XObject
   /Subtype /Form
   /BBox [0 0 100 100]
>>
stream
  /P <</MCID 5>> BDC
  ... content ...
  EMC
endstream

If this XObject is referenced on multiple pages, MCID 5 would appear multiple times.

Why This Happens:

Automated PDF generation that reuses content efficiently
Document templates with tagged repeating elements
PDF optimization that merges identical content
Export tools that don't understand PDF/UA constraints

How to Fix in Adobe Acrobat

Fixing XObject issues in Acrobat is limited because these are structural problems that may require specialized tools or document regeneration.

Detecting XObject Issues

Go to Tools > Print Production > Preflight
Search for PDF/UA profiles
Run the PDF/UA-1 compliance check
Look for XObject-related failures in results
Note which XObjects are causing issues

Using the Content Panel

Go to View > Show/Hide > Navigation Panes > Content
Expand the page content tree
Look for XObject references
Note any that appear multiple times

For Reference XObjects (30-001)

If a Reference XObject is detected:

Embed the external content: The referenced content needs to be brought into the PDF
Use Preflight fixups: Some Preflight fixups can resolve references
Regenerate the document: Often the best solution is to recreate the PDF without references
Use PDF optimization: Go to File > Save As Other > Optimized PDF and look for options to resolve references

For Duplicate Form XObjects with MCIDs (30-002)

This requires removing MCIDs from the Form XObject or making copies:

Option 1: Remove MCIDs from Form XObject

The Form XObject should become untagged (artifact)
Tag the instances separately in the tag structure
This requires understanding of PDF structure and may need specialized tools

Option 2: Create Separate XObject Instances

Instead of reusing one XObject, create copies
Each copy gets unique MCIDs
This increases file size but ensures unique identification
Requires PDF editing tools or regeneration

Option 3: Regenerate the Document

Go back to the source application
Export to PDF with different settings
Avoid settings that aggressively reuse content
Verify the new PDF passes validation

Using Preflight Fixups

In Preflight, look for fixups related to XObjects or tagged content
Some fixups may resolve the issue automatically
Test after applying fixups
Not all XObject issues have automatic fixups

How to Fix in Source Applications

Adobe InDesign

InDesign typically creates compliant XObjects:

Ensure you're using current export settings
When exporting to PDF:
- Use Adobe PDF (Interactive) or appropriate preset
- Check accessibility options
If issues persist:
- Simplify repeated graphics
- Use placed images rather than embedded objects
- Test the exported PDF

Adobe Illustrator

For graphics that will be used in PDFs:

When saving as PDF, use Save As rather than Export
Choose appropriate PDF presets
Avoid features that create Reference XObjects
Test complex illustrations in the final PDF

Desktop Publishing Software

Review export settings for PDF creation
Look for options related to:
- Object reuse
- External references
- Tagged PDF creation
Prefer settings that embed all content
Test exported PDFs for XObject issues

Avoiding Reference XObjects

To prevent Reference XObjects:

Embed all content: Don't link to external PDFs
Flatten complex documents: Convert external references to embedded content
Check export settings: Ensure "embed" rather than "link" options are selected
Avoid document assembly: Merging PDFs with references can preserve them

Avoiding Duplicate MCIDs in Form XObjects

To prevent MCID duplication:

Don't tag reusable content: Make repeated elements artifacts instead
Tag instances individually: If content must be tagged, tag each use separately
Use modern export tools: Current tools are more aware of PDF/UA constraints
Test with validators: Check PDFs before distribution

Testing Your Fix

Automated Testing

PAC (PDF Accessibility Checker):

Open the PDF in PAC
Run the PDF/UA check
Navigate to Checkpoint 30 results
Both failure conditions are machine testable
PAC will identify specific XObject issues

veraPDF:

Select PDF/UA-1 validation profile
Run validation
Look for XObject-related rule failures
Results identify specific violations

Adobe Acrobat Preflight:

Go to Tools > Print Production > Preflight
Run a PDF/UA profile
Review XObject-related failures
Check for both Reference XObjects and duplicate MCIDs

Manual Verification

After fixing:

Run automated validation to confirm issues are resolved
Check the document structure for consistency
Test with a screen reader to verify content is properly read
Ensure all content is accessible within the document

Validation Checklist

No Reference XObjects present in document
Form XObjects with MCIDs are not multiply referenced
PDF/UA validator reports no Checkpoint 30 failures
Document structure is internally consistent
All content is embedded (no external dependencies)
Repeated graphics are handled appropriately (untagged or unique instances)
Screen reader can access all content correctly

Technical Background

Understanding XObjects

XObjects are efficiency features in PDF:

Image XObjects: Store raster images
Form XObjects: Store reusable vector/mixed content
Reference XObjects: Point to content in external files

The PDF format allows XObjects to be defined once and referenced many times, reducing file size when the same content appears repeatedly.

MCIDs and Tagging

Marked Content Identifiers (MCIDs) link content in the page stream to tags in the structure tree:

Content in the page stream is wrapped with MCID markers
The structure tree references these MCIDs
This creates the mapping between visual content and semantic structure
Each MCID should appear exactly once

Why the Restrictions Exist

Reference XObjects break the "self-contained document" principle:

Accessible documents must not depend on external resources
Users need all content available without additional files
Validation requires all content be present

Duplicate MCIDs break the structure tree integrity:

The tag structure assumes unique content identification
Multiple instances of the same MCID create ambiguity
Assistive technology cannot reliably map content to structure

Additional Resources

Official Standards and Guidelines

PDF Technical Resources

Tools

PAC (PDF Accessibility Checker) - Free PDF/UA validation
veraPDF - Open-source PDF validator
Adobe Acrobat Pro - PDF analysis and editing
QPDF - Command-line PDF analysis

This documentation is based on the Matterhorn Protocol 1.02, the definitive reference for PDF/UA validation. Both XObject failure conditions are machine testable. For the most current information, consult the PDF Association and W3C WCAG guidelines.