Automated Testing of PDF Documents

When to Test

Ideally you have a process that produces an accessible PDF by ensuring that the originating document (MS Word, PowerPoint, Google Doc, Adobe InDesign) is accessible.

If you have a PDF that has been produced by someone else and it is used in a core process of your unit, it is a very good idea to test it.

Getting Started

  1. Make sure you have Adobe Acrobat DC (part of Adobe Creative Cloud). It is available for all university faculty, staff and students. 
  2. Launch Adobe Acrobat DC (Acrobat) and open the document you want to test. 

Quick Test

  1. When the document is loaded, launch the Accessibility Check option. If it is not listed in the applications pane on the right, search for it (“Accessibility”) and then choose it from the results.
  2. Keep the default Accessibility Checker Options that will appear in a dialog and click Start Checking.
  3. The results will appear in the report in the left hand panel. They will also be saved to your computer. This may be useful to share with the document creator in case they can address the issues in the source document.

For more assistance, see Adobe's documentation on its Accessibility Checker.

Interpreting the Results

You may only be interested in a pass/fail. Or you may be interested in the nature of the issues found and how these impact users with disabilities. We will address both cases below.

Pass/Fail Conditions

Any “Failed” result regarding the following will make the document inaccessible to a significant number of users:

Fail Condition Explanation

Document > Image only PDF

All of the content in the PDF is an image, including what looks like text. Many types of assistive technology will be unable to perceive the content.

Document > Tagged PDF

None of the content has semantic meaning. Headings, tables, paragraphs, lists, etc. are not programmatically identified as such. Users of screen readers will not perceive the function of different parts of the text. All of the text will be read as a long run-on sentence.

Page content > Tagged content

Same as above.

Page content > Tagged multimedia

Multimedia is not described as such

Page content > Tab order

Users who are constrained to the keyboard (and devices that emulate the keyboard) will experience serious difficulty navigating the content because the order provided to them does not match the visual order.

Forms > Tagged form fields

Form fields are not identified as such

Forms > Field descriptions

Form fields may have a visual label but they themselves are not logically labeled, so users of screen readers will not perceive what they are asking for

Alternate text > Figures alternate text

Images are lacking an alternative text that will make them meaningful to users of screen readers

Tables > Any failure

Improperly configured tables make it impossible to users of screen readers to perceive their structure

Headings > Appropriate nesting

Improperly nested headings jumble the logical order of the document

Important: There are a number of things that the scan will not be able to flag. Even if all results listed above pass, you will also need to do a manual check. See the section below. Any failure in the manual check will result in a document failure.

Manual Checks

Color Contrast

Some of your users will have a variety of low vision problems. Let’s look at the PDF through their eyes. If you spot any cases that may have poor contrast between text and background, they probably are a problem. Text over background images may be particularly problematic as you will have to compare light text over the lightest portion of the image and dark text over the darkest portion of the image.

Reading Order

If the reading order of the PDF has not been properly set, the document will be a “section salad” to a screen reader. Fortunately Acrobat has a tool to help you determine if this is the case.

  1. In Acrobat choose Reading Order from the right hand tool menu. If this is not there, search for “Reading order” in the search bar.
  2. The sections in the document will be numbered.
  3. Take a look at the numbering - does it make sense? If you were to read this document visually, would it be the same order as is indicated in the numbering? If the answer is no, this is a failure.

You Have an Inaccessible PDF. What Next?

If you have the source document, issues can be corrected there and re-exported to produce an accessible PDF.

If you do not, there are some things you can do to the PDF itself to correct it, but they are quite a bit more laborious. For help with this process, see Adobe's documentation on remediation of PDFs and PDF Remediation Basics (YouTube video, 47 minutes).