The Fox, PDFs and Screen Readers

Introduction

In nursery rhymes and fables the fox is usually the villain. In this case, though, the fox is the hero. We are discussing here Firefox’s really nice access to PDF files for screen reader users.

But I couldn’t resist a rant about the importance of accessible documents while providing some background to the sometimes vexed relationship between screen readers and PDFs.

Some background

Quite a lot of screen reader users will tell you that “pdf” and “access” should not be used in the same sentence. That is partly because the structure (or lack of) of many PDFs is a blight on humanity. In this author’s experience though, a well structured PDF makes for a very useful resource in the right circumstances. Those circumstances do not include material to be read online, where HTML is the most appropriate format.

This is not a discussion on how or why PDFs should be tagged. If you want a detailed discussion on creating accessible PDFs and remediating those that are not, see Ricky Onsman’s PDF Accessibility on the Web Part 1 and Part 2. As a quick getting-started, a correctly structured document saved as PDF from MS Office will result in a tagged PDF.

Reading a PDF with a screen reader

Let’s boldly assume for now that we are dealing with a correctly tagged PDF. There are several tools that can be used. Acrobat Reader is a strong candidate, as it understands the tag structure. Putting aside the above assumption for a moment, if the file is not tagged screen reader users should receive a message, saying that the document needs to be prepared for reading and there is a pause while that is done. One option when launching Acrobat Reader is whether to have the file presented one page at a time or as the whole document. If the file is large and whole document is chosen, the whole show can grind to a stop.

There are other options for reading PDFs with a screen reader, some free and some not. Most do not make use of the tags to the same extent as Acrobat Reader, but can make for a more nimble reading experience.

Enter the Fox

Recent releases of Firefox include support for tagged PDFs. That means that elements such as headings, lists, tables and images are presented correctly to screen readers. Unlike Acrobat Reader, there is not the risk of a large file bringing the computer to its knees. With a little familiarisation, navigation is very snappy. Jumping to a specific page is not as intuitive as in Acrobat Reader, where ctrl+shift+n allows a page number to be entered. Similarly, as far as I can tell there isn’t a keyboard command (ctrl+pgup/down in Acrobat Reader) to select previous or next page. There are Previous and Next buttons which can be reached fairly quickly with shift+b in NVDA or JAWS, but those extra keystrokes can add up.

There is one feature that Acrobat Reader and Firefox have in common. If one wants to be able to jump to specific pages, it is necessary to have only one page presented at a time. How to do this in Firefox is not immediately obvious. Select Tools Menu, arrow down to the radio button for page scrolling and check it. To reveal the whole document again, check the Vertical Scrolling radio button. While that may seem a bit clunky, it is probably faster than going into Acrobat Reader’s Preferences and making the change.

Here is something I really like. Firefox presents each page in the PDF as a separate region. With all pages available (Vertical Scrolling selected), usually pressing d with NVDA or r with JAWS (next landmark) will jump to the next page. Add the Shift key to go back a page. Some files, notably forms, may have more than one region per page, but the process is still very convenient. The usual screen reader commands for finding headings at any or a specific level are, of course, also available.

Not all plain sailing

When I first started using Firefox to read PDFs I found a couple of issues. One was that some words apparently ran together when using a screen reader. The other was that row headings in tables were not presented when moving down a column. I raised bug reports and both have been corrected in nightly builds and when Version 149 is released.

Another issue is that the Bookmarks tree, a feature I really like about PDFs, is not presented in Firefox. I have also raised that as a bug and am hoping for a fix.

PDF forms in Firefox

Unless something has changed recently, it is not possible to save a completed form when using Acrobat Reader (I do not have a copy to test) unless some extra work is undertaken when preparing the form in Acrobat. When filling out a form in Firefox, however, the form can be saved.

That’s the good news. The bad news is that field labels are not presented as nicely to screen readers in Firefox as in Acrobat Reader. I can feel another bug report coming on.

Untagged PDFs

This blog by Mark Stephens suggests that about 20% of PDFs on the internet are tagged. While the figure may be rubbery, it is a damning reflection on efforts (or lack of) to make material accessible. When various assistive technologies are being used and an untagged PDF is opened in Acrobat or even the free Acrobat Reader, the situation can be somewhat corrected. Reading is put on hold while temporary tags are applied to the file. This process can make a marked difference to the reading experience. Firefox and other tools such as Thorium Reader do not apply this remedy.

I recently received blood test results which made for interesting reading – not so much the results as the layout. In Firefox results were spread across the line one after the other. This made for, at best, tedious interpreting. When opening the file in Acrobat results were presented in nicely marked up tables. Because I have the paid (over-priced and with accessibility glitches) version of Acrobat I was able to use its automatic tagging process and save as a tagged file. When opening the new file in Firefox, it read nicely.

Most users of assistive software do not have access to Acrobat. While Acrobat Reader’s temporary tagging process is helpful, it is (a) by no means infallible and (b) must be repeated each time an untagged file is opened.

Conclusion

Firefox’s ability to present so effectively tagged PDFs to screen reader users is very pleasing. With a bit of familiarisation it makes for a very convenient interface. I do hope they can get it to present the equivalent of Bookmarks that are available in Acrobat Reader, with that lovely tree structure.

But for reading that 80% or so of PDFs that are not tagged, things are much less clearcut. Depending on content and the process used to create the file, making sense of it may or may not be possible.