SpeechDispatcher for LSR

[Peter Parente]

> I just started playing with LSR and Speech Dispatcher.

Great. Thanks for trying it out.

> I had a little problem with compilation,

Compilation meaning the install process or running LSR?

> when the pygtk.require("2.0")
> method in pyLinAcc/Atk.py stripped the correct path to the gtk-2.0
> module from sys.path and so the following import atk, gtk couldn't
> import gtk and raised an exception.

Strange. The pygtk.require call is pretty standard for all pygtk apps.
The docs at pygtk.org recommend making that call before doing the
'import gtk'. What distro/release are you using? What version of LSR?
What version of pygtk?

> The start of lsr was very smooth and it picked up Speech Dispatcher even
> without any need to tell it so (I do not have IBM TTS) and it said the
> welcome message.

Glad that worked. We're shooting for zero config at first run by
falling back through  our device definitions.

> When I run it with the developer profile, I can easily
> read the contents of all its windows and navigate their menus. However
> no other application I tried produces any sound. From the Task Monitor,
> it doesn't seem my navigating in other applications affects lsr in any
> way at all.
> I have accessibility switched on in Gnome 2.14.2. Do you have any idea
> what am I doing wrong?

Interesting. So the LSR dialogs are accessible, but nothing else. Are
events showing up in the raw event monitor? Can you use another AT
like gnopernicus or at-poke?

> Also, I'd like to ask about language switching in LSR. I see it is
> currently not supported in the Speech Dispatcher driver. Users can still
> set the default prefered language for LSR in Speech Dispatcher
> configuration. This is however not enough. The user must typically
> switch languages at runtime based on context (what language text am I
> writing? is this application localized to my language? etc.)

Language switching is supported by the LSR design but not implemented
yet in any of our devices. The style object passed to a device
includes a field for Language which can be set to any ISO
country/language code. It's up to the LSR device class to do the
switching in its respective engine, if it is supported. LSR style
objects can be sent at any time, for any data, so language switching
can be done per session, per application, per utterance, etc.

Pete
_______________________________________________
Speechd mailing list
address@hidden
http://lists.freebsoft.org/mailman/listinfo/speechd