W3C home > Mailing lists > Public > public-indie-ui@w3.org > December 2012

Re: IndieUI-ISSUE-8 (TTS context dictionary): Need a way to define a context and pronunciation dictionary on a per-resource basis. [IndieUI: User Context 1.0]

From: James Craig <jcraig@apple.com>
Date: Mon, 03 Dec 2012 12:25:55 -0800
Message-id: <9F17CF94-F525-4CE9-A23F-AFE1CB83CCB9@apple.com>
To: Independent User Interface Task Force <public-indie-ui@w3.org>
Actually, it might be better to do it this way, so that window.tts.phonetics could return a dictionary of all set values such as:
{ '✔':'True', '✘':'False', 'Knowbility':'nobility', 'WAI-ARIA':'way aria' }


// contextual text alternatives for unicode chars
// (otherwise these would be spoken literally as the unicode descriptions such as "heavy ballot x"
window.tts.phonetics['✔'] = 'True';
window.tts.phonetics['✘'] = 'False';

// contextual pronunciation of proper nouns or other terminology
window.tts.phonetics['Knowbility'] = 'nobility';
window.tts.phonetics['WAI-ARIA'] = 'way aria';


On Dec 3, 2012, at 12:16 PM, Independent User Interface Task Force Issue Tracker <sysbot+tracker@w3.org> wrote:

> IndieUI-ISSUE-8 (TTS context dictionary): Need a way to define a context and pronunciation dictionary on a per-resource basis. [IndieUI: User Context 1.0]
> 
> http://www.w3.org/WAI/IndieUI/track/issues/8
> 
> Raised by: James Craig
> On product: IndieUI: User Context 1.0
> 
> Possibly need a way to define page context dictionaries for text-to-speech. 
> See https://github.com/darobin/respec/issues/108 for some background.
> 
> Basically, you could define (in one location) that ✔ and ✘ are spoken as "True" and "False" rather than needing to define it in the markup for each use.
> 
> <span role="text" aria-label="True">✔</span>
> <span role="text" aria-label="False">✘</span>
> 
> We could also use something like this for phonetic hints for proper nouns or terminology that is uncommon outside the context of the document (and therefore mispronounce by the text-to-speech engine), where the text is not replaced, but the string is used to hint at the correct pronunciation.
> 
> I was thinking something along the lines of:
> 
> // contextual text alternatives for unicode chars
> // (otherwise these would be spoken literally as the unicode descriptions such as "heavy ballot x"
> window.accessibility.tts.phonetics.setValueForKey('✔', 'True');
> window.accessibility.tts.phonetics.setValueForKey('✘', 'False');
> 
> // contextual pronunciation of proper nouns or other terminology
> window.accessibility.tts.phonetics.setValueForKey('Knowbility', 'nobility');
> window.accessibility.tts.phonetics.setValueForKey('WAI-ARIA', 'way aria');
> 
> It's possible this should use a phonemic alphabet for the pronunciation strings, but I'm not sure about the l18n implications of that decision at the time of this writing.
> 
> 
> 
> 
Received on Monday, 3 December 2012 20:26:45 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 3 December 2012 20:26:46 GMT