Tue, 07 Aug 2007
I run a dual-seated configuration on my home computer (two monitors, two keyboards, two mice, etc.). It works for most applications, but it has one problem, which I haven't found a solution for yet: sound. I have only one 6-channel sound card, and I would like to use it from both heads (ideally split to two channels for one user, two channels for the other one, and the remaining two channels for Music Player Daemon output, connected to the stereo in the living room).
There are big problems with multiple users accessing a single sound card - usually when some application from one user's session locks the sound card (or create the ALSA DMIX shared memory segment or whatever), other apps cannot use it (which is pretty annoying especially when an incoming call through Ekiga arrives).
The solution would be to use a sound server (such as EsounD or aRts). But unfortunately not all sound apps have plugins for the same sound server. Each sound server appears to invent its own new incompatible API. However, today I have found something which looks like it can fulfill my needs in this area, and even better, it will be included in the Fedora 8 (see the feature list of F8): the project named PulseAudio.
There is an interesting presentation (PDF warning) on PulseAudio. Hopefully it will be a project of a "what people really need" nature instead of the "not invented here, let's create our own solution" one. Does anybody know whether it is possible to configure PulseAudio for two users (as described in the first paragraph)? From the examples from the PDF presentation it seems possible.