Rob Hudson <captinlogic at gmail.com> wrote: > That sounds like Pulse grabs the whole device and won't let it go. > This is why I route mplayer to another card. I'm not sure if that'll > work either, however. I don't know if Pulse grabs all audio cards in > the system or not. Pulse sure sounds evil lol. I think it does exactly that by default. You can change the configuration however. You can also run pasuspender to suspend it, and there's a command under pacmd to suspend individual cards. In general, though, I think the right way to deal with Pulse is to work with it rather than against it by sorting out whatever the underlying problems are. I think there's enough talent within the accessibility and Linux audio communities to do this, but, so far, I haven't seen a good description of what the real issues are or what needs to be done. It seems that too many people are content with work-arounds and no one is doing the real work to track down the root causes and fix them. That's a real cause for concern, because in the long run it will only mean more problems for new and existing users. Pulse is here to stay. I have it working satisfactorily on my laptop and not quite satisfactorily on my desktop system. I submitted a patch to Debian to ensure that the Espeak package was compiled properly with Pulse support (previously, the Pulse support was being overwritten during the build process). There's an open Debian bug about problems with Pulse and Emacspeak speech servers. I'm also waiting for patches to be integrated into FreeSWITCH to support Pulse - there are people working on those already. Pacmd is an interesting tool. If there is active audio input/output, you can actually get a list of all the applications that are interacting with the Pulse server, and you can adjust the volume of the audio for each application. You can also move applications from one audio device to another, though I haven't experimented with that yet. There are many other features as well. The key to making Pulse work reliably is to make sure that nothing you're using tries to bypass it by writing directly to the Alsa devices.
