W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > January to March 2004

Re: How to specify pronunciation of ABBR?

From: Charles McCathieNevile <charles@sidar.org>
Date: Sat, 10 Jan 2004 02:19:01 +0100
Cc: w3c-wai-ig@w3.org
To: Phill Jenkins <pjenkins@us.ibm.com>
Message-Id: <F8E6CF9D-430A-11D8-B212-000A958826AA@sidar.org>

This is really beyond the scope of HTML.

How to say a particular word is normally handled by Text-to-speech  
systems, and is one of the things that distinguishes good ones from  
ordinary ones. There is a specification called SSML being developed by  
the Voice group, and currently in Candidate Recommendation:  
http://www.w3.org/TR/speech-synthesis/

You could use XSLT to provide SSML styles for particular pronunciation.  
The Pronunciation Lexicon seems to be low on the priority list for the  
Voice group at the moment, but could also play a role.

(Linking a lexicon is the sort of thing that people have complained  
about as entirely impractical, or have suggested is a good way to  
approach the problem of using clear language...)

Cheers

Chaals

On Saturday, Jan 10, 2004, at 00:15 Europe/Rome, Phill Jenkins wrote:

> Quoting from the XHTML 2.0 spec
> http://www.w3.org/TR/2003/WD-xhtml2-20030506/mod-inline- 
> text.html#edef_inline-text_abbr
>
> Note that abbreviations often have idiosyncratic pronunciations. For
> example, while "IRS" and "BBC" are typically pronounced letter by  
> letter,
> "NATO" and "UNESCO" are pronounced phonetically. Still other  
> abbreviated
> forms (e.g., "URI" and "SQL") are spelled out by some people and  
> pronounced
> as words by other people. When necessary, authors should use style  
> sheets
> to specify the pronunciation of an abbreviated form.
>
> Does anyone have a example of how to specify the pronunciation with  
> CSS?
>
> I know how to specify ACSS [see note 1] for it to speak it softly or
> loudly, whether to spell it out with "speak" property set to  
> "spell-out",
> or add some sounds, etc. but not how to pronounce it.  Forget for a  
> moment
> the problem of screen readers and text-to-speech synthesizer (TTS)
> supporting it, I'm really focused on how an author should specify it so
> that the TTS speaks it as the author desires it to be pronounced.
>
> For example, how would I specify the way to pronounce the string  
> "VoiceXML"
> as "Voice X M L" and still display it without a space between the word
> voice and XML?  This might work:
>
> ABBR { speak: spell-out }
> ...some in-line text then
> Voice<abbr title="Extensible Markup Language">XML</abr> followed
> by more in-line text...
>
> Ok.  But how would I specify the correct pronunciation for non
> abbreviations and title attributes such as English words red verses  
> read vs
> reed?
>
> ABBR { speak: spell-out }
> ...some in-line text then
> <abbr title="Read the Book">RTB</abr> Club followed
> by more in-line text...
>
> Should it be red the book or reed the book, only the author might know.
> Isn't that part of the pronunciation lexicon markup [see note 2] being
> worked on by the voice browser working group?
>
>
> Note 1 Aural Style Sheets
> http://www.w3.org/TR/1998/REC-CSS2-19980512/aural.html
> Note 2 Lexicon Markup Requirements http://www.w3.org/TR/lexicon-reqs/
>
> Regards,
> Phill Jenkins,  IBM Accessibility Services
> http://www.ibm.com/able
>
>
--
Charles McCathieNevile                          Fundación Sidar
charles@sidar.org                                http://www.sidar.org
Received on Friday, 9 January 2004 20:20:35 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:13:31 UTC