Copy link to clipboard
Copied
Hi all,
I'm using NVDA to read through my PDF doc while it's open in Adobe Acrobat Pro, but NVDA skips some of the information that's already tagged (such as H2 or part of the H2 text or parts of content in a Paragraph tag). I'm not sure why this is. I've deleted and the text and re-typed it. I've deleted and re-added the tags, but the screen reader still doesn't read it. Please note that this document was originally created in Word and then converted to PDF, so I know that this could be causing some problems, but are there any solutions to this issue?
Copy link to clipboard
Copied
I deleted and re-typed the text directly in Acrobat.
By @Mikiko34403636vnga
Uh-oh. I have some bad news.
Once a PDF has been tagged for accessibility, the content can't be edited in any way. Doing so breaks the tag tree and the little yellow container boxes that are inside the tags lose their content. It just vanishes and leaves empty containers, or merges tags together, or does other strange things to the content.
So one of the first laws of accessible PDFs is: Don't Edit The Content! Includes:
All changes to the content must be done in the original source file and a new PDF re-exported.
Sometimes the file can be remediated with Acrobat's AutoTag utility, but we often find that this doesn't work because the missing content can't be recovered and retagged.
Check the tag tree again, especially on the page where the edits were made. Expand the <Tags> so that you can see the individual yellow content containers. You should see the paragraph text. If the containers are empty, the edits caused them to disappear. (The content is still visible on the page, but not in the tag tree.)
Copy link to clipboard
Copied
What is NVDA please?
Copy link to clipboard
Copied
It's a screen reader, like JAWS (Freedom Scientific). I'm trying to make the PDF accessible.
Copy link to clipboard
Copied
You should be careful with acronyms. Best is to include references, explanations, or links so that people not to understand the acronym still know what it is about. What does the NVDA forum say about the problem?
Copy link to clipboard
Copied
It's not an issue with acronyms. My issue is that the screen reader called NVDA isn't reading text in my PDF doc. My problem isn't related to acronyms. NVDA the screen reader doesn't read words in sentences and sentences in paragraphs in my PDF doc, and there aren't any acronyms in these sentences. NVDA doesn't have any information about this, hence why I posted my comment here in case someone else had encourntered this problem and perhaps could offer information about it. If not, no worries. I'll figure it out or I won't.
Copy link to clipboard
Copied
It's not an issue with acronyms.
By @Mikiko34403636vnga
The issue with acronyms is that readers won't be able to understand your request, as a Google search will point to a different direction: NVIDIA.
Copy link to clipboard
Copied
NVDA is the name of the software.
Copy link to clipboard
Copied
Non-Visual Desktop Access
Copy link to clipboard
Copied
I actually am an accessibility expert (member of the PDF/UA committee that writes the PDF/UA ISO standard) and I do know what NVDA is.
Although NVDA (and JAWS) have their problems with PDF content, NVDA is a very solid screen reader that usually doesn't miss live text content.
So I'm wondering if the problem is with your PDF. Some questions:
I've deleted and the text and re-typed it. I've deleted and re-added the tags, but the screen reader still doesn't read it.
4. Did you delete and retype the text in Acrobat? Or did you go back to Word and re-export a new PDF?
Copy link to clipboard
Copied
That's great news!
I have no idea how the content was exported from Word to a PDF since my colleague performed that action and that person is on holidays until next year. I use Windows but don't know what system my colleague has/uses.
I use an enterprise version of Adobe Acrobat Pro.
I deleted and re-typed the text directly in Acrobat. I have not used Word again for any reason while working on making the PDF accessible.
Copy link to clipboard
Copied
I deleted and re-typed the text directly in Acrobat.
By @Mikiko34403636vnga
Uh-oh. I have some bad news.
Once a PDF has been tagged for accessibility, the content can't be edited in any way. Doing so breaks the tag tree and the little yellow container boxes that are inside the tags lose their content. It just vanishes and leaves empty containers, or merges tags together, or does other strange things to the content.
So one of the first laws of accessible PDFs is: Don't Edit The Content! Includes:
All changes to the content must be done in the original source file and a new PDF re-exported.
Sometimes the file can be remediated with Acrobat's AutoTag utility, but we often find that this doesn't work because the missing content can't be recovered and retagged.
Check the tag tree again, especially on the page where the edits were made. Expand the <Tags> so that you can see the individual yellow content containers. You should see the paragraph text. If the containers are empty, the edits caused them to disappear. (The content is still visible on the page, but not in the tag tree.)
Copy link to clipboard
Copied
Thank you very much for that information. So then would a screen reader be unable to read the PDF because of the edits/changes made to tags?
Copy link to clipboard
Copied
Yes.
Screen readers like JAWS and NVDA read the tag tree and the content that's in the yellow content containers.
If the content containers are empty, then most likely you'll hear crickets (aka, nothing).
The only guaranteed solution I know is to have your colleague make the edits to the Word document when they return after the holiday, and then re-export a new PDF. Point them to the link above where the blog gives instructions on how to do that.
Best to you.
Copy link to clipboard
Copied
I am following up on this even though I was not the person who originally asked. We have had to edit a PDF (we don't have the source document). But when we do we go back into the tags tree and fix them as we know about this bug. We are still getting the issue where NVDA skips content.