speech-dispatcher and espeak

[Steve Holmes]

I am using speech-dispatcher from git or I should say
speech-dispatcher-git from AUR.  Actually, I maintain this package in
AUR and pretty much configured it to run as a system wide service in
its default form.  For that, it works great under alsa.  When I was
testing all this pulse stuff, I did configure speechd to run as a
single user and commented out the DisableAutoSpawn entry at the bottom
of the conf file.  So after all that, I was starting speech dispatcher
under pulse control; I could tell that much.  I also should mention
that I changed the AudioOutputMethod to pulse also.  I also had
problems with the environment variable still being set even though the
speechd daemon had been stopped.  For some reason, even when the
execute flag is turned off on /etc/profile.d/speechd.sh, it was still
being imported when users login.  I'm not sure why that is happening
because I thought files in profile.d had to be executible in order to
be imported like that.  So I commented out the line inside that file
and successfully had speechd running with autospawn and all.  

Frankly, I much prefer to run speechd as a central service if
possible; it just seems more stable that way.  I had too many problems
getting speechd to autospawn properly last night with all this messing
about.

Also, the distortion did not go away when I attempted the
src-best-quality deal for the resample-method.  I'll look into it
again but for now, it's nice to have a stable system again with ALSA
instead of pulse for speech dispatcher stuff.  One nice thing with
pulseaudio-gnome and friends is I get a few more sounds than I used to
get so it may be worth getting pulse going but speech dispatcher is a
bloody mess right now.

On Fri, Feb 25, 2011 at 09:36:40AM -0500, Kyle wrote:
> I just thought of this. Are you running speech-dispatcher from extra
> or the speech-dispatcher-git package from AUR? Speech-dispatcher in
> extra seems to be properly configured to run in either user or
> system mode. But I couldn't get speech-dispatcher-git to run in user
> mode and didn't try running it in system mode, although it
> constantly attempted to find its socket in /var, which is only
> supposed to apply if run system-wide. It wouldn't surprise me in
> this case to find that there are other problems with the
> speech-dispatcher-git configuration. Perhaps it's currently broken
> and will be fixed in the near future, or perhaps something has
> changed in the default configuration that I'm unsure how to fix to
> make it run correctly here.
> ~Kyle
> _______________________________________________
> Speakup mailing list
> Speakup at braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup