Using Speechd with Pulseaudio

[Steve Holmes]

I thought I would start messing with pulseaudio to see how it might
work on my Arch Linux system.  So far, I have it working after a
fashion but I can't seem to stabilize the volume so it often gets all
scratchy and or distorted and I find myself constantly having to
change the volume in gnome to clear up the noise.

Let me explain my setup so far.  At the console level, I still use
ALSA to run espeak, espeakup and obviously, speakup.  That part is
ALSA all the way and I did not install the asound.conf file for making
ALSA route through pulse.  I did this so I can be assured I will have
speech immediately aaafter booting and I can thus login with speech to
normal consoles with speakup as before.  I only use speech dispatcher
for gnome and orca.  It seems like every time I startx to get gnome
going, speech dispatcher fails with a broken pipe and I then have to
logout of gnome without any speech and kill/restart speechd in a
console just to be sure it is working first.  Then I can get back into
gnome with startx and Orca begins talking.  But now the sound is all
screwed up because pulse came up with the volume seemingly too high.
I can get into system/preferences/sound and set the volume for
speech-dispatcher in the application page down to sound better.  But
when I finish up and shutdown gnome and go back later, I have no
assurance that the volume won't stay the same.  It seems to hold for
other programs like mplayer bug sspeech-dispatcher always seems to
mess up.  I am using speechd with autospawn enabled and I have set the
AudioOutputMethod to "pulse".

Since pulse is all the rage around with speech-dispatcher developers,
I hope I can get some definitive answers on how best to configure
pulse to be stable and for speechd to come up reliably when starting
gnome.

Any takers?