E-mail List Archives
Thread: Hints for screen reader speech attributes
Number of posts in this thread: 12 (In chronological order)
From: Ofer Bartal
Date: Thu, Apr 30 2020 7:06AM
Subject: Hints for screen reader speech attributes
No previous message | Next message →
Hi,
As a website developer, I'm interested in giving hints or suggestions to
the user's screen reader on how to read certain text.
For example, I'd like the following italic to be read in a higher pitch:
The *quick* brown fox jumped over the *lazy* dog
Is there a way to do that in ARIA?
Thanks!
Ofer Bartal
From:
Subject: Re: Hints for screen reader speech attributes
← Previous message | Next message →
It isn't possible to do this with ARIA, no. You may be interested in the
work of the W3C Pronunciation Task Force though.
Explainer: Improving spoken pronunciation on the web
https://www.w3.org/TR/pronunciation-explainer/
Pronunciation gap analysis and use cases
https://www.w3.org/TR/pronunciation-gap-analysis-and-use-cases/
On 30/04/2020 14:06, Ofer Bartal wrote:
> Hi,
> As a website developer, I'm interested in giving hints or suggestions to
> the user's screen reader on how to read certain text.
>
> For example, I'd like the following italic to be read in a higher pitch:
> The *quick* brown fox jumped over the *lazy* dog
>
> Is there a way to do that in ARIA?
>
> Thanks!
> Ofer Bartal
> > > > >
--
Director @TetraLogical
From: Jonathan Avila
Date: Thu, Apr 30 2020 11:49AM
Subject: Re: Hints for screen reader speech attributes
← Previous message | Next message →
It would be better to use semantic markup and provide instructions to users on how to adjust their screen reader rather than force a user into a specific speech pattern.
Jonathan
From: Lucy GRECO
Date: Thu, Apr 30 2020 12:06PM
Subject: Re: Hints for screen reader speech attributes
← Previous message | Next message →
hello:
frankly this is not something I would ever recommend as most users would
not be comfortable changing settings and you might not know how they may
have already changed their voice. a voice is a very personal setting
and changing it makes many users uncomfortable at best. For example I
just needed to change which synthesizer I was using for a few weeks and
it made me less productive. Just the way it changed words was driving me
nuts. If you want to indicate changes i have seen some very good
implementations using hidden screen reader only text saying italic start
and italic end. but only in the case wair i really need to know that does
this work otherwise it's just na anoiince smile lucy Lucia Greco
Web Accessibility Evangelist
IST - Architecture, Platforms, and Integration
University of California, Berkeley
(510) 289-6008 skype: lucia1-greco
http://webaccess.berkeley.edu
Follow me on twitter @accessaces
On Thu, Apr 30, 2020 at 10:49 AM Jonathan Avila < = EMAIL ADDRESS REMOVED = >
wrote:
> It would be better to use semantic markup and provide instructions to
> users on how to adjust their screen reader rather than force a user into a
> specific speech pattern.
>
> Jonathan
>
>
From: Farough, David (CFP/PSC)
Date: Thu, Apr 30 2020 2:46PM
Subject: Re: Hints for screen reader speech attributes
← Previous message | Next message →
I had to laugh recently when I was testing one of our applications using NVDA.
This is an app which asks 23 questions in order to determine whether participating in various political activities is a good idea.
One of the questions was:
Section 2: Level and visibility of your current position - Question 7 of 23
With which of the following would you most closely identify?
Student (Federal Student Work Experience Program, Co-op/Internship Programs, Research Affiliate Program) radio button not checked
Operational/administrative support/junior officer (e.g., GL-MAM-7, CR-4, AS-1, PM-2, EC-2, SP-4, GL-MOC-6, BI-2) radio button not checked
Intermediate officer (e.g., EC-4, CS-2, PE-3, PM-4, FB-3, TI-4, AU-3) radio button checked
When I got to the third radio button, NVDA spoke the " AU-3" as three dollars Australian.
Who'd have thought?
From: Steve Green
Date: Thu, Apr 30 2020 2:54PM
Subject: Re: Hints for screen reader speech attributes
← Previous message | Next message →
The Slack home page used to contain a heading that included the phrase 300 Mars. JAWS used to read it as three hundred million Argentine pesos for the same reason.
Steve Green
Managing Director
Test Partners Ltd
From: Ofer Bartal
Date: Fri, May 01 2020 12:25PM
Subject: Re: Hints for screen reader speech attributes
← Previous message | Next message →
Thanks!
Using some words from that link, I searched and found this old NVDA thread:
https://github.com/nvaccess/nvda/issues/4242
And wrote some a post there hopefully to revive the discussion.
On Thu, Apr 30, 2020, 16:13 Léonie Watson < = EMAIL ADDRESS REMOVED = > wrote:
> It isn't possible to do this with ARIA, no. You may be interested in the
> work of the W3C Pronunciation Task Force though.
>
> Explainer: Improving spoken pronunciation on the web
> https://www.w3.org/TR/pronunciation-explainer/
>
> Pronunciation gap analysis and use cases
> https://www.w3.org/TR/pronunciation-gap-analysis-and-use-cases/
>
>
> On 30/04/2020 14:06, Ofer Bartal wrote:
> > Hi,
> > As a website developer, I'm interested in giving hints or suggestions to
> > the user's screen reader on how to read certain text.
> >
> > For example, I'd like the following italic to be read in a higher pitch:
> > The *quick* brown fox jumped over the *lazy* dog
> >
> > Is there a way to do that in ARIA?
> >
> > Thanks!
> > Ofer Bartal
> > > > > > > > > >
>
> --
> Director @TetraLogical
>
From: Murphy, Sean
Date: Sun, May 03 2020 5:28PM
Subject: Re: Hints for screen reader speech attributes
← Previous message | Next message →
That post is specific to NVDA. The last comment in relation to Emacspeak is available within Jaws. I do agree with the EmacSpeak concept for NVDA and other screen readers. The user of the assistive technology should control what type of information they hear and how.
In relation using CSS to control screen reader speech. The ARIA approaches they have mention is a far better approach than using CSS from my view. If I get the time, I will add to the post.
Sean
Sean Murphy | Digital System specialist (Accessibility)
Telstra Digital Channels | Digital Systems
Mobile: 0405 129 739 | Desk: (02) 9866-7917
From: Murphy, Sean
Date: Sun, May 03 2020 11:56PM
Subject: Re: Hints for screen reader speech attributes
← Previous message | Next message →
LOL, Jaws did not read it as Australian dollars in your example. Wondering why NVDA did as I didn't see the $ in front of the number.
Sean Murphy | Digital System specialist (Accessibility)
Telstra Digital Channels | Digital Systems
Mobile: 0405 129 739 | Desk: (02) 9866-7917
From: Murphy, Sean
Date: Mon, May 04 2020 1:29AM
Subject: Re: Hints for screen reader speech attributes
← Previous message | Next message →
I normally don't use these settings and I am working from memory.
Jaws, VoiceOver Mac and NVDA can announce the font attribute like italic as a word.
Jaws has a sound scheme which can change the pitch, synth (TTS) being used, etc. Not sure if NVDA without a plugin, VoiceOver for Mac does this. iOS VoiceOver and Talkback from memory does not support this capability.
The caveat to the above is if the screen reader supports the font attributes or tags in the first place.
Sean
Sean Murphy | Digital System specialist (Accessibility)
Telstra Digital Channels | Digital Systems
Mobile: 0405 129 739 | Desk: (02) 9866-7917
From: Birkir R. Gunnarsson
Date: Mon, May 04 2020 5:33AM
Subject: Re: Hints for screen reader speech attributes
← Previous message | Next message →
You should be able to use <em> or <strong> or a similar HTML text
semantic element to indicate special emphasis, however screen readers
do not support text level semantics (except Jaws and then only for
very advanced users who change voice profiles, which is probably one
in 10,000).
I've been nagging about this for years but nothing happens.
It's a fine balance between authors trying to over compensate the
screen reader experience, e.g. through unnecessary hidden text and not
being able to communicate important things to screen readers.
For instance, if you insert text such as "start of italic" you force
all screen reader users to listen to a lot of clunky verbiage.
The silly examples mentioned here are screen reader dictionaries who
try to interpret commonly used phrases to make a better experience,
and fail. With Voiceover on iOS 3m is always read as "3 meters" not,
for instance, "3 minutes".
I agree that authors need to be able to affect the screen reader
handling of certain words and patterns, and I believe this should be
doable via using the proper HTML (i.e. text level semantics), (which
should be supported) or through speech CSS, which is why I find that
ongoing work particularly interesting.
On 5/4/20, Murphy, Sean < = EMAIL ADDRESS REMOVED = > wrote:
> I normally don't use these settings and I am working from memory.
>
> Jaws, VoiceOver Mac and NVDA can announce the font attribute like italic as
> a word.
>
> Jaws has a sound scheme which can change the pitch, synth (TTS) being used,
> etc. Not sure if NVDA without a plugin, VoiceOver for Mac does this. iOS
> VoiceOver and Talkback from memory does not support this capability.
>
> The caveat to the above is if the screen reader supports the font attributes
> or tags in the first place.
>
> Sean
>
>
>
> Sean Murphy | Digital System specialist (Accessibility)
> Telstra Digital Channels | Digital Systems
> Mobile: 0405 129 739 | Desk: (02) 9866-7917
>
>
From: Steve Green
Date: Mon, May 04 2020 11:25AM
Subject: Re: Hints for screen reader speech attributes
← Previous message | No next message
It's just one of numerous heuristics that both JAWS and NVDA use. In this case NVDA interpreted a capital A followed by capital U followed by a number as being a monetary value, with AU being the code for Australian dollars. It would not do that if either of the letters was lower case. I have found this behaviour to be inconsistent, which really means that I don't fully understand the heuristic they are using.
Steve Green
Managing Director
Test Partners Ltd