PDF Accessibility 101: Avoiding Common PDF Accessibility Pitfalls

Aaron Preece

For decades now, Adobe's PDF format has been one of the premier document types if you want to create easily shareable, visually appealing, and well-structured documents. For this reason, PDFs are used to share nearly anything, from bank statements to work reports, university syllabuses, and entire books. They are absolutely everywhere.

When we consider them from an accessibility standpoint, there is good news: a properly formatted, properly tagged PDF document is generally going to be accessible with assistive technology, both screen readers and screen magnifiers. Unfortunately, even though a properly created PDF document can be incredibly accessible, there are many pitfalls to consider that can make a PDF merely difficult to read or fully and completely inaccessible.

In this article, we will be discussing the most common pitfalls you might encounter when attempting to make a PDF accessible to people who are blind or have low vision. We'll also be discussing resources you can use to ensure that your PDFs are accessible.

Pitfall 1: Image PDFs

An image PDF is a PDF document that contains a series of images instead of any real text. Essentially, you have a picture for each page that together form the document. This is the most inaccessible form of PDF. Because the document contains only images, there is no content for the screen reader to access. For a blind screen reader user to access this sort of PDF, they would need to perform optical character recognition on the PDF document as a whole to be able to read anything at all within it.

In addition, these sorts of PDFs are also quite inaccessible to screen magnification users. Image text doesn't behave like actual text. At high magnification, it may pixelate or become blurry, and text reflow and other text-aware presentation features may not work properly, making these documents much more difficult to read than accessible PDFs or other documents containing actual text.

Ideally, a PDF will be generated from an existing digital document so that it does contain proper text. It is, in fact, possible to create an OCR version of your scanned document to include in the PDF, but there are additional pitfalls with this method. These are sometimes called searchable image PDFs, and essentially, the sighted user of the document is seeing the original scanned images as they are, but underneath there is an optical character recognition layer allowing for searching the document or selecting text and so on. The problem arises when the image does not align properly with the underlying text, as optical character recognition can often be incorrect.

If you can't help but use a scanned image for your PDF, you will want to perform optical character recognition upon it. With an image PDF OCR'd and proofed, you're significantly closer to an accessible document, though there are still some pitfalls to keep in mind.

Pitfall 2: Improper Reading Order

A common issue with mostly inaccessible PDFs is the reading order of the document as presented to a screen reader. Visually, a user might be seeing the document laid out in a specific way, but the screen reader is reading things in a completely different order.

This happens most commonly with multi-column PDFs. Think of a page with two columns of text. Frequently, instead of having one column read followed by the other, if the PDF isn't properly structured, you'll have, say, a line or paragraph from one column followed by a line or paragraph from the other column, followed by text from the previous column, etc. This can make a PDF document incredibly difficult to read and understand if you're using a screen reader.

You can think of this similarly to how we described image PDFs above. PDFs can be thought of as two different layers: the visual layer, which informs how the content is visually displayed on the screen, and the semantic or logical presentation, which informs software like screen readers how they should interpret the document.

Most PDF documents start out as another format when they're created. If proper standards are followed during the creation of the source document, this frequently results in a more accessible PDF document.

Pitfall 3: Accessibility Tags

Our hypothetical PDF has actual text, avoiding the pitfall of an image PDF, and also has proper reading order, avoiding the second pitfall. However, there is a chance that our document has no indicators for important elements such as headings, lists, tables, images, and so on. Without these accessibility tags, our hypothetical PDF looks like one giant block of text without any sort of differentiation, even if it is in the correct order.

It is important to ensure that the accessibility tags are properly present. This will allow the PDF to properly display things like headings, links, images, lists, tables, and so on so that the screen reader user can navigate them.

A good way to think about this is that your PDF should have the same level of accessibility as a web page would, as oftentimes, a screen reader is actually reading a PDF similarly to how it would read a web page. Just like you would use proper heading tags on a website to indicate a heading level one, instead of just changing the font, you would do the same with a source document that would become an eventual PDF.

If proper accessibility tags are included, we now have our accessible PDF.

Caveat: PDF Reader Interpretation

Unfortunately, it is possible to design a document perfectly with proper text, proper reading order, and proper accessibility tagging, and a user still might find it difficult to read based on the PDF reader they are using.

PDF readers or interpreters vary widely in how well they respect accessibility tagging. For example, the official Adobe PDF reader is quite good at respecting accessibility tags. Other PDF readers can vary widely in how well they support this functionality. In addition, people frequently use other types of software to view PDFs where displaying PDFs is not the primary function of the software. This might include web browsers, email clients, or other types of e-reader software. As an example, your PDF might render perfectly in the official Adobe Reader but not display any accessibility tagging to a screen reader at all if you open it in a browser.

Compare this to HTML documents. HTML renders very slightly differently across browsers and devices, but for the most part, if a site is structured in one way in one browser, viewing it on other browsers or other devices will display it similarly with minor differences. The same is true for a Word document. Generally, though less so than HTML documents, a Word document will render similarly no matter what word processor or viewer you're using to display it. For this reason, my personal opinion is that HTML is the ideal format for displaying content, as it's consistent across devices.

Creating an Accessible PDF Document

Now that we're aware of the pitfalls of PDF accessibility, how do we ensure that our PDFs are accessible? Ideally, the PDF will be made accessible by default and not require remediation.

A layperson-friendly way to do this is by creating the PDF using Microsoft Word, as it has extensive accessibility options that result in an accessible PDF with a high degree of accuracy. If you use Word's built-in tools for, say, creating proper headings, lists, tables, adding alt text to images, properly formatting information, etc., Word's export-to-PDF function can accurately apply the structure and tags to the PDF. Note that “Document structure tags for accessibility” should be enabled so that Word can attempt to create an accessible PDF.

In addition, Microsoft Word includes an automatic accessibility checker that will give you a list of accessibility errors that can be helpful for catching any items that might have escaped notice. That being said, accessibility checkers are helpful but aren't perfectly accurate, so it's important to be sure that you're following good accessibility practices from the start and, if possible, have a blind or low vision person check the resulting document with their screen reader to ensure accessibility, though this latter step isn't always easily doable.

Once the document is exported to PDF format, you can also use Acrobat Pro's accessibility checker, as well as the software's cloud-based auto-tagging system. It is also possible to use Acrobat Pro's reading order and tagging tools to manually remediate any accessibility issues if automated fixes or checks aren't catching them.

Note as well that there is a standard for PDF accessibility called PDF/UA. You can learn more about the standard here.

Resources

There are many, many, many excellent resources online providing step-by-step instructions and thorough documentation of methods for creating accessible PDF documents. Below, find a list of several that you might find particularly helpful.

The following includes official Microsoft and Adobe guidance on creating accessible PDF documents, information from the W3C, United States Section 508, and the PDF Association:

Author
Aaron Preece
Article Topic
Accessibility Tips and Tricks