Re: [orca-list] orca speech uncontrolled and jumbled with speech-dispatcher

["Lukas Loehrer" <listaddr1 gmx net>]

Krishnakant writes ("[orca-list] orca speech uncontrolled and jumbled with speech-dispatcher"):
> I navigated through the menus or even orca preferences for that matter.
> But the speech comes jumbled up and is uncontrolled.
> For example if I move a bit fast between the applications menu, then
> accessaries, games and graphics is said all together.
> As if some children in school are making noise together.
> I think the problem is that when output is sent to speech synthesizer by
> orca, it is forgetting to interrupt the previous chunk before speaking
> the content at locus of focus.

The problem could be at a lower layer as well. It sounds to me as if
some component of the audio stack is doing software mixing with too
large an output buffer. The problem occurs when one of the mixed
channels wants to be cancelled, the already mixed protion of its
output cannot be revoked, because it has already been mixed into the
output buffer to together with the output of the other channels that
do not want to have their output discarded. I have experienced such
problems with alsa and dmix. The solution was to set the output vuffer
to a reasonably small value.

I guess orca on Ubuntu 9.10 uses pulseaudio with in turns uses alsa
for audio output. In this setup, mixing is probably performed by
pulseaudio. I do not know how to influence mixing parameters their or
even if it can suffer from the problem I described above . In alsa,
you can use a ~/.asoundrc file or /etc/asound.conf. An example file
that works well for some peaple can be found here:

http://emacspeak.googlecode.com/svn/trunk/servers/linux-outloud/asoundrc

Best regards, Lukas