Speech-Dispatcher and SMP question

Speech-Dispatcher and SMP question

[Keith Hinton]

How well does Speech-Dispatcher currently work on SMP systems, using the 
native ESpeak module v.s a Generic module?
I know an advantage of the native module is that ESpeak won't have to be 
called and that "wav" files won't be generated for each key-press.
Just curious how Diispatcher andnative module work on SMP.
I think if anyone uses the generic one, though, that you'll still be bale to 
make dispatcher crash.
I was able to do that with the Generic dectalk module.
I was not able to do so with the Flite module however, sinse that was 
native.
But FLite is too slow to respond to keyboared stuff for me. ESPeak should be 
much better.
Another question: does the native module allow multiple sounds to go through 
the soundcard?
Regards, --Keith. 

Re: Speech-Dispatcher and SMP question

[Lukas Loehrer]

Keith Hinton writes ("Speech-Dispatcher and SMP question"):
> How well does Speech-Dispatcher currently work on SMP systems, using the 
> native ESpeak module v.s a Generic module?

I tried to make the native espeak module thread-safe when developing
it. However, I have never had an SMP system to test it and their can
always be bugs of course, especially when dealing with concurrency.

> But FLite is too slow to respond to keyboared stuff for me. ESPeak should be 
> much better.

flite is very fast. I guess the problem in speech-dispatcher is that
there is some silence produced by flite at the beginning and end of
each utterence. I once made a patch for this but never checked it into
speech-dispatcher cvs. I will try to do so next week.

> Another question: does the native module allow multiple sounds to go through 
> the soundcard?

The native espeak module can use alsa for sound output. Thus, if the other
programs are using alsa as well, they should be able to use the
soundcard simultaniously.

Best regards, Lukas

Re: Speech-Dispatcher and SMP question

[Lorenzo Taylor]

> How well does Speech-Dispatcher currently work on SMP systems, using
> the native ESpeak module v.s a Generic module?
> 
> Quite well so far. I have been running it since last night with
> absolutely no trouble. The only problem I have is that for some reason
> the Ubuntu init script doesn't work so I have to start it manually to
> get it working. But once it starts it runs like a dream.
> 
> Another question: does the native module allow multiple sounds to go
> through the soundcard?
> 
> If you have pulseaudio on your system, you can set eSpeak to play
> through it. In your /etc/speech-dispatcher/modules/espeak.conf,
> uncomment the line that says
> 
> EspeakAudioOutputMethod "alsa"
> 
> and change "alsa" to "pulse". It works pretty well here, although it
> slows down response time just a little. I still find it quite usable
> though.
> 
> If you would rather use alsa/dmix instead of pulseaudio, you don't
> have to change anything in speech-dispatcher, just configure your
> asound.conf to use the dmix plugin.
> 
> HTH,
> Lorenzo
-- 
Make a change. Make a difference.
Obama for America: our time has come!
http://www.barackobama.com
>