[LAD] IM like GUI for netjack
lunarshuttle at gmail.com
Sun Nov 30 22:43:37 UTC 2008
torbenh at gmx.de wrote:
> now that the backend is mostly done, i am thinking about
> a GUI for netjack.
> i already found gtknetsource.py on my HD, i started mucking
> with that some months ago.
> i plan to extend that thingy now.
> but setting up the connection, and getting IP addresses
> of users is still a PITA.
> i am thinking along the lines of an IM like thing,
> based on jabber.
> i am not seeing good options to making this available
> in the various IM clients.
> thats why i would rather like to have modified jabberd running
> on jackaudio.org or at the consortium servers.
> it would not support chatting or stuff.
> only show who is online. and if a session is running.
> it would only make the IP of a user available when he agrees,
> to open session... blabla... security.
> so basically you click on your buddy, to open a session.
> buddy agrees.
> IP of buddy is transmitted. tool measures connection.
> provides you with some options.
> ie compression ratio, number of channels, latency, who is master ?
> and starts the netjack session.
> thoughts ?
This sounds like things that me and Robert J fantasized about a year
ago, or probably more. We did some experiments playing 1 guitar and 1
vocal channel each, being in our respective houses (1-2 km:s away from
each other, <2ms ping) into a bad midi background version of AC/DC:s
TNT, which came from the jack master, where MusE was used to do all the
routing. It worked ok with respect to latency, however alsa-in and
alsa-out (mainly alsa-in) was pretty unstable and crashed quite a lot.
It appeared that it worked quite ok when being on the same wired LAN but
when using my WLAN these things showed up again. It appeared as if it
should be ok for us to run a session w 4 ppl realtime, but the alsa-in
crashes made it difficult to run long enough to verify that it was
actually possible since there was always someone that didn't have it up
I made a not successful attempt on trying to change alsa-in to not try
to stretch when a timeout occured, I got the feeling that there was some
kind of attempt of trying to adapt the buffer/reading to the timeouts
and that this might have been the cause of the crashes. Has anyone else
tried anything similar - perhaps there has been improvements in these
areas, I'm far from updated. The pitch-shifting stretch made our guitars
sound quite a bit strange at times as well...
So.. this sounds really cool. Via the jack homepage you could find
someone to play with, start a band, record an album, have a live gig and
split the band - all during a time period of one night. But only with
ppl living in a very short ping-time from yourself.
Anyway, one thing I can think of is that it would be nice to be able to
specify a script that is run when a client connects (if you're master -
then your favourite seq app could be told to add another input and
connect all the other ppl:s outputs to it - so you get to hear what the
other ones are playing but not have your own input echoed back) or when
you connect to a master.
Give me/us a yell if you have something that is testable.
More information about the Linux-audio-dev