On Mon, Sep 19, 2011 at 3:20 PM, Alan McKinnon <alan.mckin...@gmail.com> wrote: > On Mon, 19 Sep 2011 14:02:39 -0400 > Michael Mol <mike...@gmail.com> wrote: >> On Mon, Sep 19, 2011 at 1:42 PM, Volker Armin Hemmann >> <volkerar...@googlemail.com> wrote: >> > Am Montag 19 September 2011, 12:37:16 schrieb Michael Mol: >> > >> >> I recall reading about dmix in LinuxJournal years ago, but I don't >> >> think I ever got around to setting it up; >> > >> > you don't set it up. It just works. If your sound card does not do >> > hardware mixing (onboard sound doesn't) you are using dmix. >> >> Ah. As I said, I hadn't poked or researched dmix since I read about it >> in LinuxJournal. Pretty sure that particular issue came out over ten >> years ago. >> >> That doesn't quite jive with my experience with apps some apps >> managing to take exclusive control over sound devices. In particular, >> if, e.g. Flash were run under Firefox before WINE or PulseAudio, then >> the latter two didn't get to play.* > > Flash isn't a good example though. It just assumes that it is the most > important (only?) thing in the universe, and tries to take over the > hardware for itself. If I read recent blogs correctly, even Windows > users suffer from the same thing with Flash.
Audio, I don't *think* so. At the very least, Vista and 7 allow you to configure whether or not applications are allowed to take exclusive control over a device. Video inputs, yes. > > I think the presumption in this thread in that sound apps make *some* > attempt to play nicely - Flash doesn't fit that category. The only > category it fits is "useless crap that should either be deleted or only > used when absolutely necessary" That actually makes for a really good argument to use something like PA's ALSA wrapper when you can't do without Flash. I hear recent versions of Flash support PA directly. I can see an argument for Flash wanting control over A/V hardware; audio and video recorders have been implemented in it. Flash, IME, doesn't grab A/V until a Flash applet access them, but it also doesn't let them go. Perhaps their internal VM is poorly defined such that it's OK for apps to assume that once they have a resource, it's always there, and they're stuck maintaining that VM model for compatibility. -- :wq
