Re: Thoughts on speech

[Willie Walker <William Walker Sun COM>]

Hi Steve:

The IDL for the main speaker interface is here (it's pretty short/simple):

http://svn.gnome.org/svn/gnome-speech/trunk/idl/GNOME_Speech_Speaker.idl

The main driver interface, which is basically what you discover when 
using Bonobo, is here:

http://svn.gnome.org/svn/gnome-speech/trunk/idl/GNOME_Speech_SynthesisDriver.idl

"Extra" stuff, such as pitch, rate, punctuation, etc., is buried in the 
parameter information.  You need to hunt around to find it, and it's all 
handled by convention versus a real spec.  An example of a more complete 
set of parameters can be found at the end of this file:

http://svn.gnome.org/svn/gnome-speech/trunk/drivers/viavoice/viavoicespeaker.c

Hope this helps!

Will

Steve Lee wrote:
> Brilliant Will, I've just leant a whole lot, thanks
> What is the main API for speech? Something like speak( sText ) ?
>
> Steve
>
> On 07/03/2008, Willie Walker <William Walker sun com> wrote:
> > OK - well....let's see.  I'll fill in what I know, but I need the Speech
> >  Dispatcher folks to fill in what they know.
> >
> >  GNOME Speech:
> >
> >    Just a thin wrapper over a TTS engine.
> >
> >    Doesn't do audio management - leaves that to the TTS engine.
> >
> >    Drivers for Festival, FreeTTS, DECtalk, IBMTTS/ViaVoice,
> >    Loquendo, eSpeak, Cesptral/Swift, Eloquence, and even a
> >    wrapper for SpeechDispatcher.  No support for DECtalk
> >    Express.
> >
> >    At a minimum, callbacks supported at the utterance level,
> >    where an utterance is the chunk of text tossed at it via
> >    a single speak command.  Callbacks are also supported at
> >    the word progress level if the engine supports it.
> >
> >    Mostly just sends text off to the speech synthesis engine
> >    for speaking.  The only 'extra' stuff that's really done
> >    is adding index marks to text strings to be notified of
> >    speech progress at the word level for those TTS engines
> >    that support it.
> >
> >    No real support for SSML.
> >
> >    Audio is controlled by the speech synthesis engine.
> >
> >    Bonobo/CORBA based, essentially locking it to GNOME
> >    for all intents and purposes.
> >
> >    Speech services are discoverable and activatable as
> >    system services (via Bonobo Activation).
> >
> >    Those skilled in the art and with knowledge of the TTS
> >    engine's API can write a driver in a day.  It's much more
> >    difficult for those not skilled in the art.  ;-)
> >
> >    Difficult to debug.
> >
> >  Will
> >
> >
> >  David Bolter wrote:
> >  > Will,
> >  >
> >  > That sounds very reasonable to me.  Can you start it?  :)
> >  >
> >  > cheers,
> >  > D
> >  > Willie Walker wrote:
> >  >> Hi All:
> >  >>
> >  >> This speech issue is obviously one filled with passion and high
> >  >> expectations.  I think our ultimate end goal here is to find a
> >  >> solution that works well and fits within the various constraints.
> >  >>
> >  >> The two solutions we've been talking about, gnome-speech and Speech
> >  >> Dispatcher, both have their strengths and weaknesses, and I'm not sure
> >  >> we all understand what they are.  Nor do I think we all understand
> >  >> what "works well" means and what the constraints are.
> >  >>
> >  >> As an exercise, what do you all think of us having a somewhat
> >  >> impassioned and pragmatic discussion about the various features and
> >  >> the current state of gnome-speech and Speech Dispatcher?
> >  >>
> >  >> Will
> >  >> _______________________________________________
> >  >> gnome-accessibility-list mailing list
> >  >> gnome-accessibility-list gnome org
> >  >> http://mail.gnome.org/mailman/listinfo/gnome-accessibility-list
> >  >>
> >  >
> >
> >  _______________________________________________
> >  gnome-accessibility-list mailing list
> >  gnome-accessibility-list gnome org
> >  http://mail.gnome.org/mailman/listinfo/gnome-accessibility-list