Re: [orca-list] Patch: Experimental eSpeak support using python-espeak

[Peter Vágner <pvdeejay gmail com>]

Hello,
Well unfortunatelly nothing I have done applies to orca vs 
speech-dispatcher communication. So if I have mentioned possible fixes 
in this conversation these are just fixes to this experimental direct 
eSpeak backend. These are things I haven't understood correctly on my 
first try or I have implemented incorectly.
If you are asking about comparison to the speech-dispatcher backend and 
if you are curious what motivated me to try doing this then I would say
I don't consider this release quality at least not yet. It creates some 
performance issues when comparing to the speech-dispatcher backend. 
There are issues with interupts not being responsive enough to be usefull.
Originally I assumed we should be able to at least slightly boost 
responsiveness of the whole chain orca vs speech-dispatcher vs eSpeak by 
taking speech-dispatcher out of the chain for good. However initial 
testing shows this is either not the case or it needs additional work.
I need to test it more and try to come up with more improvements to find 
out whether this will be at least comparable to speech-dispatcher backend.
Of course when we are considering adding eSpeak specific features such 
as variants support, rate boost and perhaps there is something else too 
we can either do it by first contributing this to speech-dispatcher and 
then adapting orca to new speech-dispatcher API. I don't feel brave 
enough to implement something usefull for speech-dispatcher so I'm 
experimenting this way. eSpeak specific featues may or may not come 
later from me depending on the fact if I can make this performant enough.


Greetings

Peter

On 24.08.2015 at 15:15 Joanmarie Diggs wrote:
> Hi Peter.
>
> Can you please explain the reasons why these sorts of features and fixes
> cannot be added to speech-dispatcher?
>
> --joanie
>
> On 08/23/2015 06:25 PM, Peter Vágner wrote:
> > Hello,
> > Today I was playing with orca a bit more than usual. The result is that
> > I have implemented an alternative speech server subclass. I have taken
> > speechdispatcherfactory module and I have replaced python-speechd API
> > with python-espeak one accross the whole module. So by doing this I
> > think I have created experimental direct eSpeak support for orca. It
> > requires python-espeak to be installed.
> > I tryed to do this because a while ago we were discussing this approach
> > here on the list. You guys were hoping for additional responsiveness
> > increase and easier way on how to add eSpeak specific features like
> > variants support into orca. Also I have come up with another little
> > advantage as I have been thinking about this more. By default eSpeak is
> > built with both pulseaudio and portaudio sound outputs. It first tryes
> > to use pulseaudio and if that is not available it plays its audio
> > through portaudio thus using alsa directly. It can't switch between this
> > dynamically at runtime but it at least works out what to use at
> > initialization what translates to when launching the application when
> > using python-espeak.
> > There are no eSpeak specific features yet. Things which are working
> > include speaking, changing voices, changing pitch, rate and volume. Say
> > all support with proper progress tracking is also working.
> > I am not sure I like this better than using eSpeak through
> > speech-dispatcher like we are doing for years. I am using this
> > experiment for just a few hours now. I am attaching a patch for those of
> > you who are not afraid of risking a bit. Please make sure you know how
> > to revert this and troubleshoot your system when it fails to speak. This
> > is my first attempt at trying to implement something into orca so it is
> > very likelly it may have loads of issues I even can't think of at the
> > moment.
> > So how to apply it if you really want to test it out.
> > - Install python-espeak. If you are on arch linux, get it from the AUR.
> > I guess on Debian or Ubuntu you can just do apt-get install
> > python-espeak. With other distributions I don't know whether there are
> > prebuilt packages. Python-espeak is hosted on launchpad and if there are
> > no packages for your distro then you can get it from there and install
> > using python setup tools.
> > To check whether the patch applies cleanly into your cloned orca git
> > tree you can run
> > git apply --check espeakfactory.patch
> > If there are no conflicts found you can then apply it by doing
> > git apply espeakfactory.patch
> > After doing this you can just build and install orca like normal.
> >
> > Greetings
> >
> > Peter
> >
> >
> >
> > _______________________________________________
> > orca-list mailing list
> > orca-list gnome org
> > https://mail.gnome.org/mailman/listinfo/orca-list
> > Visit http://live.gnome.org/Orca for more information on Orca.
> > The manual is at http://library.gnome.org/users/gnome-access-guide/nightly/ats-2.html
> > The FAQ is at http://live.gnome.org/Orca/FrequentlyAskedQuestions
> > Log bugs and feature requests at http://bugzilla.gnome.org
> > Find out how to help at http://live.gnome.org/Orca/HowCanIHelp