Re: [math-on-web] meeting minutes, 2017/03/30

Mark Hakkinen mhakkinen@ets.org<mailto:mhakkinen@ets.org> at ETS has done a lot with SSML and has also successfully made modifications to Alexa using SSML to pronounce some Finnish family names correctly.

Thanks
EOM

Charles LaPierre
Technical Lead, DIAGRAM and Born Accessible
E-mail: charlesl@benetech.org<mailto:charlesl@benetech.org>
Twitter: @CLaPierreA11Y
Skype: charles_lapierre
Phone: 650-600-3301



On Apr 3, 2017, at 2:51 AM, Peter Krautzberger <peter.krautzberger@mathjax.org<mailto:peter.krautzberger@mathjax.org>> wrote:

Hi everyone,

Below are the minutes from the last meeting.

Best,
Peter.


# Math on web CG, 2017/03/30

* Present: Jean, Volker, Peter, John
  * regret: Charles, Daniel, Kevin, Moritz

* Volker: could we have calendar invite?
  * Peter: wasn't sure that worked but Kevin proved me wrong
  * => ACTION Peter.
* Peter: regrets for lack of minutes from previous meeting
  * only Daniel and me, chatted about layout and didn't take notes
  * nobody showed up so we lost track of taking minutes => should pick up on this in the next meeting
* Peter: minutes for monday a11y tf meeting, I missed
  * ACTION Daniel promised to share his notes
* Peter: finished previous action => posted https://www.peterkrautzberger.org/0192/
  * current results, more work ongoing
  * will have to redo some of the recordings (even I can't remember what i did :-( )
  * overall quite happy with the results so far
* Volker: web speech API experiments
  * SSML doesn't seem supported in real life
  * Question: has anyone else experience with this?
  * Peter: do you have a sample that could be forked?
    * Volker: yes, will post.
* Peter: in other news, the A11Y OM https://github.com/WICG/aom/
  * what Houdini is for the paint ree, this is for the a11y tree
* Volker: quick sample for dev tools
  ```
  synth = speechSynthesis; ll = new SpeechSynthesisUtterance('<SPEAK> x <PROSODY RATE="-5%"> equals </PROSODY> <BREAK TIME="200ms"/> start frac <PROSODY PITCH="+15%"> <PROSODY RATE="-5%"> minus </PROSODY> b <PROSODY RATE="-5%"> plus minus </PROSODY> Square root of <PROSODY RATE="+10%"> b </PROSODY> </PROSODY> <PROSODY RATE="+10%"> <PROSODY PITCH="+32.5%"> square </PROSODY> <PROSODY PITCH="+15%"> <BREAK TIME="300ms"/> </PROSODY> </PROSODY> <PROSODY PITCH="+15%"> <PROSODY RATE="+5%"> minus </PROSODY> <PROSODY RATE="+10%"> four a c </PROSODY> <BREAK TIME="400ms"/> </PROSODY> over <PROSODY PITCH="-15%"> two a </PROSODY> <BREAK TIME="400ms"/> end frac </SPEAK>');
  ll.voice = synth.getVoices()[1];
  synth.speak(ll);
   ```
   * theoretically works, but no browser seems to accept prosody
   * the SSML worked on the IBM Watson demo, https://text-to-speech-demo.mybluemix.net/
   * student project tried it with Alexa and that worked
   * unfortunately, this crashes browsers and even
   * working on a related demo using  festival engine with sable markup
   * Peter: where's the speech api spec?
     * can't find group, just DVCS access to spec
  * [more discussion and hacking]
  * Jean: it's hard to find out implementation support
  * Peter: for Mac I found https://stackoverflow.com/questions/21952736/the-right-way-to-use-ssml-with-web-speech-api#comment46115096_21952736
  * https://developer.apple.com/library/content/documentation/UserExperience/Conceptual/SpeechSynthesisProgrammingGuide/FineTuning/FineTuning.html#//apple_ref/doc/uid/TP40004365-CH5-SW3
  * `synth = speechSynthesis; ll = new SpeechSynthesisUtterance('the square root of [[pbas +4]] 2 [[char LTRL]]a[[char NORM]] to the [[pbas +4]] 14 [[char LTRL]]x[[char NORM]]'); ll.voice = synth.getVoices()[1]; synth.speak(ll);`
  * Jean: tried, seemed to have an effect.
  * ACTION: Volker to share a demo (for real this time ;-) )