[LAU] Jack vs Jack2, PulseAudio, Speech, Commandline!

Len Ovens len at ovenwerks.net
Sun May 26 05:56:30 UTC 2013


On Thu, May 23, 2013 5:45 pm, Janina Sajka wrote:
> Len Ovens writes:
>> I thought about this for a bit and I am not sure if you mean that two
>> users can be logged in at the same time and both listen to sound (PA
>> would
>> have to run system wide) or just that only one user would login at a
>> time
>> but you want the same setting for all in which case you only need to
>> change the system settings.
>>
> Actually, at the same time. I tend to have 23 consoles open plus the one
> dedicated to the graphical desktop. Sounds like overkill, but there's
> method to my madness, i.e. gui on tty1, mail on tty10, my work with the
> W3C on tty7,
> etc., etc.

I didn't know you could do so many. I wonder what you then do with the
GUI. I don't know what pulse needs to work on a terminal like that. I have
gotten jackd working across several (6 or 7) VTs/ssh terms by using a
script to tie the dbus instance to each login. But they are all one user.
I know pa creates a pseudo alsa device for alsa enabled apps to connect to
and that may not need dbus, but I think apps connecting direct to PA, do
need dbus.

>> --
> I typically have 1 or 2 ttys with root login for system level things.
> Yes, I know sudo, but some commands actually won't work without a native
> root login.
>
> And, I have a user that builds rpms.
>
> All of these need screen reader support, and that works just fine over
> alsa, but no so well over pa--except that the Speech-Dispatcher, Espeak,
> plus Orca combo actively supports pa.

It sounds like you have at least 3 users, so system wide might work best.

Another option that may work just as well is to use "screen" just under
the screen reader. Screen (in case you haven't used it) allows multiple
ttys as well, but because it is only one app you only need one instance of
your screen reader. The tty's under it can be logged in either with ssh
root at localhost or sudo bash --login. They may be a root bash, but the
screen reader would work at the user level that screen is running at.
Actually, screen says it supports multi-users natively.

> Card 0 for Speakup with TTSynth and whatever incidental sound
> Card 1 is a digital out of 0, I don't use it.
>
> Card 2 for Orca. This one gets pa
>
> Card3 is my Sennheiser headset for SIP calls, preferably FreeSwitch, but
> that's horribly latent at the moment.
>
> Card4 is currently my HDSP Multiface. Needless to say, this one's for
> jack--once I get it working!
>
> Card5 is an old Chaintech PCI card that uses ICE1725 drivers, and has
> pretty good quality. I use it with mplayer, routing its output to some
> old JBL Ions. If I continue to have problems with the Multiface, this is
> my fallback until I can get something better.
>
> So, I'm OK to turn pa on/off for 2 and 5, but definitely not 0 or 4, and
> probably not 3. This, plus the multi logins.
>
> The UDEV rule does seem promissing. I just need to learn more about how
> to identify devices to UDEV. And, since Cards 2 & 3 are USB, I should
> arguably lock them in via UDEV, rather than counting on
> /etc/modprobe.d/local.conf alsa assignments to stay the same.
>
> Probably more than you needed to know about my setup!

It doesn't sound impossible... Using VTs just makes it harder :) I found
the best way was to get dbus going on one and save the environment (at
least the dbus parts to a file. Then have each new VT export that to their
environment as they were logged in. If you are doing a GUI, then use the
dbus stuff from there for any VTS logged in as the same user.

-- 
Len Ovens
www.OvenWerks.net



More information about the Linux-audio-user mailing list