5 Replies Latest reply on Dec 1, 2013 6:31 PM by CtDave

    Read Out Loud won't read tags, reads content instead

    DejanTeranet

      Hi,

       

      I have a PDF generated through a third party library, and it is a tagged PDF (running it in PAC 1.3 and 2.0 says it passes accessibility test) and when I use the Read Out Loud feature on it in Adobe Acrobat Pro XI and Adobe Reader, it ignores all the tags (including any Alternate and Actual text fields) and reads the content of the PDF instead of the tags. It even reads the header and the footer which are marked as Artifacts, which is strange as Artifacts should never be read out loud. When I run NVDA and have it read the PDF, it reads it normally as expected (reads only the tags, and follows the Alternate/Actual text fields).

       

      http://www.sendspace.com/file/4ouw92

       

      Here's an example of the file. I would expect the Read Out Loud feature to start with "I don't know", but instead it says DEVELOPMENT (an artifact) and then I D N, then I expect it to get to the last line of text before the footer and then simply continue to the second page, but it will instead read the header and the footer before it moves on to the next page. Any reason why it behaves like this? Why it ignores the tags and why it doesn't ignore the artifacts?

       

      Thanks,

       

      Dejan

        • 1. Re: Read Out Loud won't read tags, reads content instead
          CtDave Level 5

          Dejan,

           

          ROL is not an assitive technology feature. It is a simple, simple text to speech feature. Adobe has never presented it as anything other than that.

          Being what it is ROL has no "awareness" of PDF tags or logical strucure.

           

          Be well...

          1 person found this helpful
          • 2. Re: Read Out Loud won't read tags, reads content instead
            DejanTeranet Level 1

            Hi Dave,

             

            Thanks for you quick reply. I understand that the ROL feature is not A.T. and I also read other replies by you on this subject to other people, but if I run it on a simpler PDF that was built fully in Adobe Acrobat, it reads the tags properly, like in this PDF:

             

            http://www.sendspace.com/file/ta2bky

             

            It does not read the Header and the "Page 1 of 1" footer, and it reads the Actual text field for the paragraph tag, thus FYI is read out loud as For Your Information. So it seems that ROL feature is aware of the tags, but for some reason it ignores them in the PDF I create programatically using a third party library, so I am wondering if that means there is something wrong with the pdf.

             

            Thanks,

             

            Dejan

            • 3. Re: Read Out Loud won't read tags, reads content instead
              a C student Level 3

              Hi Dejan,

               

              When I view your first example file in the Acrobat content pane, I notice some of the Artifact containers are nested inside of content containers. I suspect that is part of the problem. But - when I try to fix things I get Acrobat to crash, or ROL to say "Warning empty page", or several other crazy errors. I tried a few other things like converting to Word and printing to PDF and XPS - all with odd unhelpful results. The file seems to be corrupt or possessed or it just does not like me. Also, I was not able to duplicate your success with PAC 1.3 or 2. In my experience the example file fails both. Sorry to not be helpful. I think you are on the right track by suspecting is wrong with the PDF - the file itself may be broken.

               

              a 'C' student

              1 person found this helpful
              • 4. Re: Read Out Loud won't read tags, reads content instead
                DejanTeranet Level 1

                Hi "a C student",

                 

                Thank you for your reply.

                 

                In order to make a sample of the pdf I changed actual code to produce the PDF I have linked to in my original post, and I didn't actually check that it passes PAC 1.3 and 2.0, as I simply believed that my code change did exactly the same thing that the production code does. It of course did not, and I gave you a "possessed" PDF Sorry about that. Here is a new sample, this time done a bit better

                 

                http://www.sendspace.com/file/ndajik

                 

                In PAC 1.3 it only warns about bookmarks, in 2.0 it fails the font embedding which should not be an issue and says PDF/UA identifier is not set - this might be the issue, but I do not know how to set it, I understood that all I need is to have the Title, Language and some other metadata fields set for the PDF to be PDF/UA compliant.

                 

                Anyway, I hope this sample works out better for you

                 

                Thanks,

                 

                Dejan

                • 5. Re: Read Out Loud won't read tags, reads content instead
                  CtDave Level 5

                  I understood that all I need is to have the Title, Language and some other metadata fields set for the PDF to be PDF/UA compliant.

                   

                  Sorry. No. A PDF, to be PDF/UA compliant requires rather more.

                  Do read / study ISO 14289-1 and 14.8 of ISO 32000-1

                   

                  Some additional informative reference information:

                  http://www.aiim.org/Research-and-Publications/Standards/Committees/PDFUA 

                   

                   

                  Be well...