WebAIM - Web Accessibility In Mind

E-mail List Archives

Thread: PDF Text Reading Weird w/ Screen Reader

for

Number of posts in this thread: 4 (In chronological order)

From: Carly Gerard
Date: Tue, Sep 26 2017 9:39AM
Subject: PDF Text Reading Weird w/ Screen Reader
No previous message | Next message →

Hello WebAIM,


I'm working on making an accessible PDF, but after adjusting tags and running the checker, my coworker and I found the text to read weirdly, like letters in words would separate from each other. For example, one phrase might visually say "Click your name in upper right hand corner and select 'Manage Your Account Information- from dropdown menu..." However, when using VoiceOver (my coworker used Window Eyes), the text might actually sound like:


"Click your name in uppe r right hand c orner and s elect 'Manage Your Account Information- from dropdown m enu..."


I am not sure how to apply all the standard accessibility features in this PDF without the text reading this way. Any thoughts or suggestions are appreciated!


Carly

--


Carly Gerard | Web Accessibility Developer
= EMAIL ADDRESS REMOVED = <mailto: = EMAIL ADDRESS REMOVED = >
University Relations & Marketing
Western Washington University

From: Philip Kiff
Date: Thu, Sep 28 2017 6:15AM
Subject: Re: PDF Text Reading Weird w/ Screen Reader
← Previous message | Next message →

On 2017-09-26 11:39 AM, Carly Gerard wrote:
> [....] I'm working on making an accessible PDF, but after adjusting tags and running the checker, my coworker and I found the text to read weirdly, like letters in words would separate from each other. For example, "Click your name in uppe r right hand c orner and s elect 'Manage Your Account Information- from dropdown m enu..."
My guess is that this is caused by extraneous p tags being inserted in
the middle of words by whatever software generated the PDF to being
with. The screen reader will usually pause at the end of each p tag.
This can be fixed by using the "Touch Up Reading Order" tool in Acrobat
Pro by simply selecting the troublesome words or groups of words as a
single phrase all at once and then clicking "Text".

If that doesn't work, then my next guess would be that the actual
content contains unwanted spaces or hidden paths or containers that
somehow got stuck in between the letters in the content tree. This can
happen if your text has underlining, or shading, for example. You can
try to fix this by viewing the content manually and cutting and pasting
or dragging individual letters or words into a single container in the
order you want the text to be read with the correct spaces in the
correct order. You may have to re-tag the text using the Touch Up
Reading Order tool after you do this. This kind of manual adjustment of
the actual content can be finnicky and tricky, and changes are often
non-reversible, so be careful to save regularly if you attempt to make
such changes.

Phil.

From: Metzessible
Date: Thu, Sep 28 2017 1:19PM
Subject: Re: PDF Text Reading Weird w/ Screen Reader
← Previous message | Next message →

Hi Carly,

There could be any number of reasons why the PDF isn't working right, but
it's most likely how the PDF engine interpreted the original authoring
content as it was exported from whatever original file format into PDF. In
any case, that's not likely to be any help to you now that your PDF is
almost finished.

There's a number of different methods that you could do to try to address
this situation, all varying in degree of complexity. You could try to use
the Touch Up (TORU) tool to just manually select all the text, but that can
really affect how the tag structure is laid out. There's only a handful of
the most common tags available in here, and if you have a really well
structured document (as you should), then it might change something else.
Also, there's a chance it won't do any good because there are in fact empty
spaces in there. In general, I tend to avoid the TORU tool like the plague
unless I absolutely have to mess with content (like tables). Anyway, Here's
the steps I would take:


1. Save the PDF as a new version (Just In Case)!
2. Double click the word that is causing you issues to highlight it.
3. In the Tag options menu, select "Find Tag from Selection..."
4. The tag pane should open the tags and navigate directly to the
content you just highlighted. (Sometimes this step needs to be repeated due
to a bug in Acrobat, but it usually happens after the second time.)
5. You should see the Parent tag (I'll assume it's <P>). Directly under
that is what we affectionately call a bankers box that contains real
content.( This is the marked content that has a direct association with the
content on the page).
6. With the bankers box highlighted, go back to the word that is causing
you issues and select it again. Notice the content changes in color from
blue to gray, but both are now highlighted.
7. Right click the gray bankers box and choose, "Create Tag from
Selection...".
8. The text you just highlighted is now inside the bankers box. You
might need to reorder the bankers boxes inside the Parent tag so it will
read correctly (IMPORTANT: Don't drag and drop the text inside an existing
bankers box or it'll get deleted. You'll need to repeat step 6 above if you
do this).
9. Now that you've reordered it correctly, right click the new bankers
box that has the word or phrase you just created and select New Tag...
10. In the New Tag dialog box, use the drop down menu and select 'Span'
(you can also type the tag name, but it must be spelled correctly using the
proper capitalization and spelling as it would appear in the Tag tree. The
way they appear in the dropdown menu are the human-language equivalents of
the tag names).
11. Move the newly created bankers box under the newly created <Span>
tag and make sure it's still in the right place in the content flow under
the parent tag.
12. Right click <Span>, and choose 'Properties' (it'll be at the bottom
of the context menu).
13. In the Tag Properties dialog box, navigate to the Actual Text field.
Actual Text works like replacement text, and incidentally it's what we use
for OCRed documents. Basically anything that gets put in here will read
like text instead of what's actually there.
14. Type the word exactly as it should read ("upper", "right" "corner"
"select", etc if you are only doing words, or 'Manage Your Account
Information- from the dropdown menu." if you chose a phrase, etc.).
NOTE: There's a weird bug in JAWS where it will concatenate the Actual Text
with the following word in the parent tag (haven't tested this with other
SRs). I tend to add a physical space after each word so it will read
correctly.
15. Try that out and let us know if it reads correctly. If not, then it
might be a content issue, which is a bit more advanced, but I can help you
sort that out.

To visualize:

Steps 1-7:

<P>
--[] Click your name in the upper right hand corner and...
--[] 'Manage Your Account Information- from dropdown...

(* You might not see any physical difference in the tag tree, and my
opening and closing brackets represent bankers boxes)

Step 8

<P>
--[] Click your name in the
--[] right hand corner and select...
--[] upper
--[] 'Manage Your Account Information- from dropdown menu"

"Click your name in uppe r right hand c orner and s elect 'Manage Your
Account Information- from dropdown m enu..."

Step 8 (fixed)

<P>
--[] Click your name in the
--[] upper
--[] right hand corner and select...
--[] 'Manage Your Account Information- from dropdown menu"

Step 9+

<P>
--[] Click your name in the
--<Span>
----[] upper
--<Span>
----[] right hand corner and select...

Hope this helps! Please let me know if this works. Feel free to reach me
off list if you need a hand with this bad rocket.

Jon

On Thu, Sep 28, 2017 at 8:15 AM, Philip Kiff < = EMAIL ADDRESS REMOVED = > wrote:

> On 2017-09-26 11:39 AM, Carly Gerard wrote:
>
>> [....] I'm working on making an accessible PDF, but after adjusting tags
>> and running the checker, my coworker and I found the text to read weirdly,
>> like letters in words would separate from each other. For example, "Click
>> your name in uppe r right hand c orner and s elect 'Manage Your Account
>> Information- from dropdown m enu..."
>>
> My guess is that this is caused by extraneous p tags being inserted in the
> middle of words by whatever software generated the PDF to being with. The
> screen reader will usually pause at the end of each p tag. This can be
> fixed by using the "Touch Up Reading Order" tool in Acrobat Pro by simply
> selecting the troublesome words or groups of words as a single phrase all
> at once and then clicking "Text".
>
> If that doesn't work, then my next guess would be that the actual content
> contains unwanted spaces or hidden paths or containers that somehow got
> stuck in between the letters in the content tree. This can happen if your
> text has underlining, or shading, for example. You can try to fix this by
> viewing the content manually and cutting and pasting or dragging individual
> letters or words into a single container in the order you want the text to
> be read with the correct spaces in the correct order. You may have to
> re-tag the text using the Touch Up Reading Order tool after you do this.
> This kind of manual adjustment of the actual content can be finnicky and
> tricky, and changes are often non-reversible, so be careful to save
> regularly if you attempt to make such changes.
>
> Phil.
>
> > > > >

From: Carly Gerard
Date: Thu, Sep 28 2017 2:17PM
Subject: Re: PDF Text Reading Weird w/ Screen Reader
← Previous message | No next message

Hi Jon,


Thanks so much for replying! I'll be sure to keep this explanation in mind for future PDFs if I encounter the issue again.


I ended up trying the PDF again by exporting to a plain text version (no formatting or any other gunk that might throw the text off), applied basic styles in Word, and then tagged the PDF. According to Adobe AC and my colleague using JAWS, it looks and sounds much better! It might've been going from .odf to .docx to .pdf causing some conversion issues.


I do like your solution and hope to use it in case a file gets wonky, so that I don't have to constantly redo them!

From: WebAIM-Forum < = EMAIL ADDRESS REMOVED = > on behalf of Metzessible < = EMAIL ADDRESS REMOVED = >
Sent: Thursday, September 28, 2017 12:19:26 PM
To: WebAIM Discussion List
Subject: Re: [WebAIM] PDF Text Reading Weird w/ Screen Reader

Hi Carly,

There could be any number of reasons why the PDF isn't working right, but
it's most likely how the PDF engine interpreted the original authoring
content as it was exported from whatever original file format into PDF. In
any case, that's not likely to be any help to you now that your PDF is
almost finished.

There's a number of different methods that you could do to try to address
this situation, all varying in degree of complexity. You could try to use
the Touch Up (TORU) tool to just manually select all the text, but that can
really affect how the tag structure is laid out. There's only a handful of
the most common tags available in here, and if you have a really well
structured document (as you should), then it might change something else.
Also, there's a chance it won't do any good because there are in fact empty
spaces in there. In general, I tend to avoid the TORU tool like the plague
unless I absolutely have to mess with content (like tables). Anyway, Here's
the steps I would take:


1. Save the PDF as a new version (Just In Case)!
2. Double click the word that is causing you issues to highlight it.
3. In the Tag options menu, select "Find Tag from Selection..."
4. The tag pane should open the tags and navigate directly to the
content you just highlighted. (Sometimes this step needs to be repeated due
to a bug in Acrobat, but it usually happens after the second time.)
5. You should see the Parent tag (I'll assume it's <P>). Directly under
that is what we affectionately call a bankers box that contains real
content.( This is the marked content that has a direct association with the
content on the page).
6. With the bankers box highlighted, go back to the word that is causing
you issues and select it again. Notice the content changes in color from
blue to gray, but both are now highlighted.
7. Right click the gray bankers box and choose, "Create Tag from
Selection...".
8. The text you just highlighted is now inside the bankers box. You
might need to reorder the bankers boxes inside the Parent tag so it will
read correctly (IMPORTANT: Don't drag and drop the text inside an existing
bankers box or it'll get deleted. You'll need to repeat step 6 above if you
do this).
9. Now that you've reordered it correctly, right click the new bankers
box that has the word or phrase you just created and select New Tag...
10. In the New Tag dialog box, use the drop down menu and select 'Span'
(you can also type the tag name, but it must be spelled correctly using the
proper capitalization and spelling as it would appear in the Tag tree. The
way they appear in the dropdown menu are the human-language equivalents of
the tag names).
11. Move the newly created bankers box under the newly created <Span>
tag and make sure it's still in the right place in the content flow under
the parent tag.
12. Right click <Span>, and choose 'Properties' (it'll be at the bottom
of the context menu).
13. In the Tag Properties dialog box, navigate to the Actual Text field.
Actual Text works like replacement text, and incidentally it's what we use
for OCRed documents. Basically anything that gets put in here will read
like text instead of what's actually there.
14. Type the word exactly as it should read ("upper", "right" "corner"
"select", etc if you are only doing words, or 'Manage Your Account
Information- from the dropdown menu." if you chose a phrase, etc.).
NOTE: There's a weird bug in JAWS where it will concatenate the Actual Text
with the following word in the parent tag (haven't tested this with other
SRs). I tend to add a physical space after each word so it will read
correctly.
15. Try that out and let us know if it reads correctly. If not, then it
might be a content issue, which is a bit more advanced, but I can help you
sort that out.

To visualize:

Steps 1-7:

<P>
--[] Click your name in the upper right hand corner and...
--[] 'Manage Your Account Information- from dropdown...

(* You might not see any physical difference in the tag tree, and my
opening and closing brackets represent bankers boxes)

Step 8

<P>
--[] Click your name in the
--[] right hand corner and select...
--[] upper
--[] 'Manage Your Account Information- from dropdown menu"

"Click your name in uppe r right hand c orner and s elect 'Manage Your
Account Information- from dropdown m enu..."

Step 8 (fixed)

<P>
--[] Click your name in the
--[] upper
--[] right hand corner and select...
--[] 'Manage Your Account Information- from dropdown menu"

Step 9+

<P>
--[] Click your name in the
--<Span>
----[] upper
--<Span>
----[] right hand corner and select...

Hope this helps! Please let me know if this works. Feel free to reach me
off list if you need a hand with this bad rocket.

Jon

On Thu, Sep 28, 2017 at 8:15 AM, Philip Kiff < = EMAIL ADDRESS REMOVED = > wrote:

> On 2017-09-26 11:39 AM, Carly Gerard wrote:
>
>> [....] I'm working on making an accessible PDF, but after adjusting tags
>> and running the checker, my coworker and I found the text to read weirdly,
>> like letters in words would separate from each other. For example, "Click
>> your name in uppe r right hand c orner and s elect 'Manage Your Account
>> Information- from dropdown m enu..."
>>
> My guess is that this is caused by extraneous p tags being inserted in the
> middle of words by whatever software generated the PDF to being with. The
> screen reader will usually pause at the end of each p tag. This can be
> fixed by using the "Touch Up Reading Order" tool in Acrobat Pro by simply
> selecting the troublesome words or groups of words as a single phrase all
> at once and then clicking "Text".
>
> If that doesn't work, then my next guess would be that the actual content
> contains unwanted spaces or hidden paths or containers that somehow got
> stuck in between the letters in the content tree. This can happen if your
> text has underlining, or shading, for example. You can try to fix this by
> viewing the content manually and cutting and pasting or dragging individual
> letters or words into a single container in the order you want the text to
> be read with the correct spaces in the correct order. You may have to
> re-tag the text using the Touch Up Reading Order tool after you do this.
> This kind of manual adjustment of the actual content can be finnicky and
> tricky, and changes are often non-reversible, so be careful to save
> regularly if you attempt to make such changes.
>
> Phil.
>
> > > > >