Screen reader access to PDF: a users guide

This guide is intended for screen reader users who want to learn about practical techniques to access PDF.

While the keyboard commands included apply to JAWS version 7, the main principles will be useful for any screen reader user.

There are twelve sections in this guide.

1. Get the right version of Adobe Reader

Adobe Reader is the free software you can download to open and read PDF. Older versions of Adobe Reader were called Acrobat Reader.

Adobe recommend upgrading to the latest version of Adobe Reader to take advantage of new accessibility features.

Only the full version of Adobe Reader (around 35 megabyte) supports accessibility features. Be aware that other versions such as Adobe Reader standard and Adobe Elements do not provide the same features.

Full version accessibility features include Accessibility Quick Check, text reflow and Read Out Loud. Read out loud is the text-to-speech engine built into Adobe Reader.

When installing Adobe Reader on your computer for the first time, the presence of assistive technology on your system triggers the install shield to include five extra pages of options. The default options are recommended.

Download the latest copy of Adobe Reader.

2. Configuring Adobe Reader to work with assistive technology

Adobe have published guidelines on their website explaining how to access PDF documents with assistive technology.

Adobe recommend assistive technology users familiarise themselves with these guidelines.

3. Opening PDFs on web pages

PDFs displayed in a browser may not be fully compatible with screen readers. The same goes for Microsoft Word documents and Excel spreadsheets when viewed inside a browser. This is because the JAWS scripts that are loaded are optimised for browsing HTML rather than other file types.

It is therefore recommended that all web-based PDF documents are opened in Adobe Reader not within your browser.

To set Adobe Reader to open in its own window for PDF documents on the web follow these three steps:

• Open Adobe Reader, then open the "Document" menu, and then select "Accessibility Setup Assistant".
• Proceed through each dialogue box and click "Next" until you get to the fifth screen.
• On screen five the very last option is called "Display PDF Documents In the Web Browser". You should make sure this option is not checked.

Adobe publish the above guidance together with other guidance on customising Adobe to suit you on their Adobe Reader 9 accessibility pages.

4. Saving PDF's from web pages

You can save a downloadable file directly onto your hard drive from a web page using the 'Save target As' feature. Knowing about this function is important if you want to get the PDF converted using Adobe's conversion services (section 6 below) or to send the PDF to another recipient such as the original service provider (section 11 below).

To operate this function you'll need to follow these 2 steps:

• Activate the right mouse click when focus is on the download link. Either move the JAWS cursor to the link then press the right mouse button key on the NumPad or press Shift and F10 when the PC cursor is focused on the download link.
• A context menu will pop up, select "Save target as…" and then use the usual file location dialogue box to save the file to your preferred location.

5. Tips on navigating PDFs

Standard reading keys

If you are using JAWS, you can use the standard reading commands or Navigation Quick Keys to navigate and read PDF documents in Adobe Reader.

• Cursor down - reads next line
• Cursor up - read previous line
• Cursor right - read character to the right

Extract from JAWS Help "Press INSERT+H to view a complete list of JAWS hot keys for HTML or PDF Documents".

Generally, you can use these same keystrokes. However, we've found the keystrokes don't always have the same effect.

Paragraphs

The usual paragraph navigation keys (Control and cursor up or down) used to move through paragraphs in MS Word documents and web pages may not work. We've found this keystroke can fail to move through paragraphs and instead skips whole blocks of text.

Similarly, the JAWS Quick Nav Key to navigate paragraphs (the "P" key when Quick Nav Keys is On) also fails to move through paragraphs and instead skips blocs of text.

When paragraph navigation keys fail to work correctly, you'll have to resort to navigating line by line using cursor keys. This is obviously slower.

Furthermore, unlike MS Word documents or web pages, JAWS does not speak the blank lines that separate paragraphs. As such it isn't even possible to tell when a new paragraph starts or ends. We can find no way round this problem.

Headings

You may already be using the JAWS Navigation Quick Key "H" to navigate by headings in a web page. This functionality was first built into JAWS version 6. JAWS version 7 and above now extends Navigation Quick Keys to Microsoft Word and Adobe Reader. Refer to notes below as not all headings are recognised.

Navigation Quick keys are preset to "On" within web pages and Adobe Reader, but "Off" for MS Word. To switch Nav Keys On and Off follow one of these two options:

• Press "Insert" and "V" to open the JAWS Verbosity dialogue box and use Spacebar to toggle Nav Quick Keys to "On" and hit the "Return" key to close. Or,
• Use the short cut key which is "Insert" and "Z" which toggles the setting on and off without needing to open the Verbosity dialogue.

Available Nav Quick Keys range across most of the alphabetic keys, "P" jumps to next paragraph, "H" to next heading, "l" to next list and t to next table. Hold "Shift" down with any Nav key to navigate backwards. When Quick Keys are On, press "Insert" and "1" to switch Keyboard Help On letting you explore what each Quick Nav key does.

JAWS can also display lists of certain elements in the document. Extract from JAWS Help "press INSERT+F7 for a list of links, INSERT+F6 for a list of headings, or INSERT+F5 for a list of form fields. In addition, if you hold down CTRL+INSERT while pressing a Navigation Quick Key, JAWS displays a list of all elements of that type on the page".

Note, for these list features and the Nav Quick keys to work, the headings must be correctly formatted. In the case of web pages the headings must be H1, H2 etc. for MS Word documents they must be set as Heading 1, Heading 2 etc and applied using Word Styles. For PDF, they must be tagged headings.

Tables

To navigate tables hold down "Control" and "Alt" then use the cursor keys. This function only works when the table is tagged, either by the original author or when Adobe Reader assigns temporary tags.

Non-uniform tables where cells are merged may not have a logical reading order. It is not possible to know as a JAWS user whether the non-uniform structure affects the accessibility of the table. A human reader assistant who can check the visual appearance is required.

If you find you cannot navigate beyond a certain point this is because a block of merged cells is restricting the cursor movement. Try navigating along until the underlying table structure does allow movement beyond that column or row.

Note, where table layouts have been created using unstructured techniques such as spaces, tabs or text boxes, these will not appear as navigable tables to JAWS users.

Hyperlinks

Use "Tab" and "shift" "Tab" to navigate links on the page. The "Return" key activates links.

Form input fields

Use Tab and shift Tab to navigate form input fields on the page. Extract from JAWS Help system "If you are entering information into a form field, be sure to switch to Forms mode first by pressing ENTER. To exit Forms mode, press NUM PAD PLUS".

Note, PDF forms may be slower to respond to commands than in other file formats.

Pages and bookmarks

Navigate pages by pressing "Control" and "Page up" or "Page down" keys. This is different from navigating pages inside MS Word or a browser.

The Navigation pane in Adobe Reader allows you to browse and navigate the whole document by the available bookmarks or page numbers. Press F6 to switch focus into this pane.

Within the Navigation pane use the "Tab" key to move between fields and cursor keys to browse the lists. You can use the navigation pane to select and navigate bookmarks and page numbers.

Pressing the "Return" key on any listed bookmark or page number should switch focus to that point in the document and switch focus back out of the Navigation pane. The focus remains on the Navigation pane.

Use F6 to move the focus away from the Navigation Pane and back to the document pane. JAWS doesn't announce when it's open or closed and we can find no way of setting any announcement.

Notes:

• Navigation pane is a non-standard dialogue box and the fields and controls can be tedious to operate with JAWS.
• Bookmarks mark the headings in a tagged PDF.
• Page numbers are organised by odds and evens if the document is a two-page spread design, use right and left arrows to move horizontally between the columns of odds and evens.

6. Converting PDFs to other formats

There are a number of ways to convert PDF into alternative formats. The options are increasing as new software products come onto the market. This is a snapshot at time of writing. Note, automated converter tools may not always successfully convert PDFs. Essential information may be missing or in the wrong reading order depending on how the original document was designed. Adobe provide two conversion services:

• Online conversion service - lets you convert Adobe PDF files to text or HTML, that are in English and most West European languages. For more information, visit Online conversion tools for Adobe PDF documents.
• Email conversion service - lets you attach a PDF document to an email and the system returns a converted document in the specified format. Send your documents in an attachment to pdf2txt@adobe.com or pdf2html@adobe.com.

You can also convert PDFs into other formats using the following software applications:

• Adobe Reader - this free software allows PDFs to be saved into a text only (txt) file. The output is an unstructured and unformatted text file which does not support page elements such as headings, table layouts or hyperlinks.
• Adobe Acrobat Standard and Professional - allows the creation editing and conversion of PDF to Word, doc and other file types. It costs from £350 to £750. You can find out more about the Acrobat family of products on Adobe's website.
• ScanSoft PDF Converter - lets you convert PDF to an MS Word or RTF document. It costs $50. Find out more on the ScanSoft website.
• ScanSoft PDF Converter Professional - lets you create, edit and convert PDFs. It also allows conversion of PDF files to audio files using Nuance's RealSpeak Text-To-Speech technology. It costs $100. Find out more on ScanSoft website.

Note, we have not tested Nuance's RealSpeak Text-To-Speech output. Information on their website seems to indicate the speech output will be generated as an unstructured MP3 stream.

7. Understanding PDF

Unlike word processed documents that comprise text, graphics and formatting assembled in the order they appear on screen, a PDF file is a free form database.

There are three types of PDF:

• Image only - this most simple form of PDF contains only graphics and the coordinates to mark them up on the page. A paper document that has been scanned into PDF format is most likely to be an image only PDF. These are totally inaccessible to screen readers.
• Searchable image - these are similar to above, but contain a library of text keywords linked to co-ordinates that allow users to search and jump to sections of the document. Again, the text contained is unsuitable as a means of accessing the document by screen reader.
• Text and formatting - can contain text, formatting, tables, spreadsheets, video and animations, user input fields, 3D graphics etc. This type of PDF can contain an accessibility layer.

8. Understanding Adobe Reader

When a PDF file opens using Adobe Reader, the program redraws the graphical information held in the database precisely on the page using coordinates. The advantage is the visual appearance of the original document is preserved very accurately. The disadvantage is Adobe Reader may not be able to generate equivalent quality information in the accessibility layer. That is to say if the original author of the PDF did not optimise the PDF for accessibility.

Most PDFs available today are not tagged and are not optimised for accessibility. This is because while many hundreds of software applications can output PDFs, only a few are capable of producing tagged PDF or provide tools to optimise PDFs for accessibility.

When you open a PDF document that is not tagged, Adobe Reader analyzes the document and automatically adds tags to improve the reading order. These tags are only temporary, and must be added each time the PDF document is opened.

Be aware many PDFs do not even contain text and formatting. If the PDF you have is 'image only' or 'Searchable image' then Adobe Reader cannot generate any text in which to place in the accessibility layer. The only way to check is either to wait for Adobe Reader to warn you that the PDF contains no information or use the Quick Accessibility Check from the Document menu. These reports are only a rough guide though.

Extract from Adobe's website "Adobe Reader automatically analyzes all untagged PDF documents and produces temporary tags that describe the document structure to assistive technology."

Be aware many PDFs do not even contain text and formatting. If the PDF you have is 'image only' or 'Searchable image' then Adobe Reader cannot generate any text in which to place in the accessibility layer. The only way to check is either to wait for Adobe Reader to warn you that the PDF contains no information or use the Quick Accessibility Check from the Document menu. These reports are only a rough guide though. Extract from Adobe's website "Adobe Reader automatically analyzes all untagged PDF documents and produces temporary tags that describe the document structure to assistive technology."

9. Reading untagged PDF using Adobe Reader

When Adobe Reader reports that the PDF you are trying to open is untagged, this means there is no information in the accessibility layer. Adobe Reader will attempt to convert what information it can from the PDF and place it in the accessibility layer.

Below we describe what we've found Adobe Reader is able to translate into the accessibility layer and what it seems to miss.

• Navigational elements such as headings and links are not tagged. The processed document is therefore not properly structured. The JAWS "H" quick nav key will not work and it therefore isn't possible to identify heading text or use them to navigate.
• Table layouts are tagged so it is possible to use "Control" plus "Alt" plus arrow keys to navigate tables, but note the column and row headers may not be aligned correctly and any merged cells within the table layout may render the reading order invalid. It isn't possible to check this without a sighted assistant.
• Graphical elements such as pictures, special symbols and charts are tagged as a graphic, but no text labels are generated it therefore isn't possible to check whether essential information is missing.
• Reading order is most likely to be correct when the document being processed is generated from MS Word or Corel Word Perfect. You can check which application generated the PDF by selecting "Properties" option from the "File" menu. The short cut key is "Control" and "D".
• Reading order is least likely to be correct when the PDF was produced from graphic design or page layout software, such as Quark XPress. Most publishing teams and graphic designers use Quark in the UK. Again you can find out which application generated the PDF by selecting "Properties" option from the "File" menu. The short cut key is "Control" and "D".

There is no reliable way for a screen reader user to check whether the reading order of information in the accessibility layer matches that shown in the visual layer. The only guaranteed method is to use a sighted assistant to check the reading order using one of the following methods:

• Save the PDF as text only (using the File menu "Save As Text" option described above) then compare the plain text version with the PDF version, edit and correct the text file where problems are found.
• Audition the reading order by using the "Read Out Loud" feature. Check reading order on-screen matches that spoken by the speech output. Where problems are found, generate a text only file and edit as needed.

Note, auditioning the reading order can be a very awkward method and not suited to long documents.

10. Reading tagged PDF using Adobe Reader

Adobe say: "Best results can be achieved, however, if the file has been optimized by adding "tags" to the file. Tags contain information about document structure and substantially increase accessibility. They store information such as header locations, hyperlinks, and alternative text descriptions for graphics. Assistive technology interprets these tags and displays the information on a variety of platforms, from screen readers to refreshable Braille output devices."

Significant issues we've found with tagged PDF:

• Any content that was missed or that was incorrectly tagged will not be accessible. Feedbak indicates that many document producers may find Acrobat Professional difficult to use and so there's a risk that authors fail to tag content correctly.
• The Accessibility Quick Check feature found on the Document menu can only report whether there are tags present in the PDF. It cannot indicate whether the tagging is accurate, meaningful or if tags are missing.
• The only way to check the validity of tagging is to submit the PDF for a full accessibility check by an experienced Acrobat Professional operator who is skilled at tagging PDFs.

11. Your rights under UK Disability Discrimination Act (DDA)

The UK Disability Discrimination Act 1995 (DDA) makes it unlawful for service providers to pass on the cost of making information accessible to disabled people.

Making PDFs accessible or providing them in an alternate format is likely to be a reasonable adjustment under this act.

As a disabled end user you have a right to request that the information supplied to you in PDF is accessible.

When you receive a PDF or attempt to download a PDF from a website we recommend:

1. Check that the information provider has provided a fully accessible PDF that complies with accessibility standards. If Adobe Reader reports that 'This document is untagged and must be prepared for reading' then you cannot be sure of the accessibility of the document. It is recommended to contact the provider to obtain a fully accessible version of the PDF or an alternative means of access.
2. Where the service provider fails to make the necessary accessibility adjustments in a reasonable time, you may have to consider the following courses of action:
   • Complain to the service provider stating that you have a right of access to the information. You can find out more about your rights by following the link to Your Rights on the Disability Rights Commission UK website.
   • Seek assistance from someone who is skilled in making PDFs accessible. This does mean you have to carry the cost of making the adjustments yourself, but many users say they cannot wait weeks to read a document so are forced to do this.

12. JAWS standard reading keys

You can use the standard JAWS reading commands or Navigation Quick Keys to navigate and read PDF documents.

• Go Back a page, ALT+LEFT ARROW
• Next link, TAB
• Prior link, SHIFT+TAB
• Open a link, ENTER
• Display a list of links, JAWSKey+f7
• Display and manage PlaceMarkers, control+shift+k
• Display a list of form fields, JAWSKey+f5
• Display a list of headings, JAWSKey+f6
• Display a list of frames, JAWSKey+f9

To move through a web page:

• Next Anchor, A
• Next Button, B
• Next ComboBox, C
• Next different element, D
• Next Edit, E
• Next Form Field, F
• Next Graphic, G
• Next Heading, H
• Next List Item, I
• Jump to a specific line with J
• And return to the starting point before the jump with Shift+J
• Next Place Marker, K
• Next List, L
• Next Frame, M
• Skip past links, N
• Next Object, O
• Next Paragraph, P
• Next Block Quote, Q
• Next RadioButton, R
• Next Same Element, S
• Next Table, T
• Next unvisited link, U
• Next visited link, V
• Next Checkbox, X
• Next Division, Z
• Step past element, > (greater than)
• Step prior to element, < (less than)
• Add SHIFT to these commands to move to the prior item
• Add Control+JAWSKey to the applicable quick keys to bring up a list of that element
• For example, Control+JAWSKey+z will bring up a list of the divisions on the page

To move and read in tables:

• Next row, WINDOWS KEY+DOWN ARROW
• Prior row, WINDOWS KEY+UP ARROW
• Read Row, WINDOWS KEY+COMMA
• Read Column, WINDOWS KEY+PERIOD
• Next cell in a row, ALT+CTRL+RIGHT ARROW
• Prior cell in a row, ALT+CTRL+LEFT ARROW
• Cell below in a column, ALT+CTRL+DOWN ARROW
• Cell above in a column, ALT+CTRL+Up ARROW
• Jump to cell,Control+ j

End of guide.