=?iso-8859-1?Q?Re:_[Kde-accessibility]_KTTSD_does_not_speak, _no_way_to_debug?=

[webmaster at gebetgenerator.de]

Hi!

Gary Cramblitt <garycramblitt at comcast.net> schrieb am 31.07.2004,
01:18:18:
> Why it doesn't work for you, but does work for me and others is a
> mystery.

The QMutex used in order to syncronize between the two threads is not
totally thread-safe, as Gunnar just told me. The effect is that the
first sentence gets spoken, but kttsd does not return to the idle
state. However, this doesn't occur every time.

> 1.  Plugins no longer need to audibilize the speech. They need only 
> create a wave file from the text. KTTSD will take care of sending the
> file to aRts (or other player plugins such as gstreamer). This allows
> an utterance to be synthesized simultaneously while another is
> audibilized.

I am still wondering how we can integrate the possibility to generate
wave files to the cross-platform low-level API we where discussing. I
plan to talk about this at aKademy with some people knowing
gnome-speech. A KDE multimedia guy (Scott) recommended using portaudio
within our low-level API, as there are bridges to almost every
multimedia streaming API for Linux and other systems. However, this
wouldn't allow the pre-generation of wave files.

Of course we should make the generation of wave files an option, so
portaudio could still be used where wave files are not wanted - but we
would need a cross-platform API for the generation of wave files as
well.

> Gunnar, do you mind if I modify the Hadifax plugin?  If you mind, I
> can create an alternate version called mbrola.

Gunnar and I are currently on holiday till August 4. Let's discuss the
changes after we return to Germany.

Olaf