[pulseaudio-discuss] a sounding program with several output streams

[Roman Beslik]

On 05.01.12 12:53, Colin Guthrie wrote:
> 'Twas brillig, and Roman Beslik at 05/01/12 00:24 did gyre and gimble:
>> On 03.01.12 14:37, Colin Guthrie wrote:
>>> 'Twas brillig, and Roman Beslik at 02/01/12 22:55 did gyre and gimble:
>>>> them) for enhanced user experience. E.g., Skype outputs incoming call
>>>> ringing to speakers and talk to a handset. Because programs do not chose
>>> Yes, programs can create as many streams as they want (each sink will
>>> only support (I think) 32 streams, but that's just to prevent too much
>>> craziness! :)
>> Yeah, that's good. Actually incoming call ringing is not visible because
>> it is included into "System Sounds", I was perplexed and read other
>> perplexed users' discussions, and decided that there is some defect in
>> PulseAudio or Skype.
> Yeah, this is intentional. Event sounds, like a buddy signing in, are
> very short lived, so as was mentioned in your other thread, they are
> simply grouped up under "Event Sounds" such that the user does actually
> have a real chance of adjusting their volume to suit their tastes. I
> appreciate it might not be as transparent as it could be however.
> Perhaps there is something we can do about that in the GUIs. Such as
> show the names+icons of the 5 most recent applications that have
> produced an event sound, just so that it's a bit more obvious to users
> looking to adjust the volume.
That would be fine.

>>> Over all this metadata allows for interesting things to happen
>>> automatically. For example if someone has a headset (USB or Bluetooth)
>>> we can automatically use it for phone calls which is (with ~90%
>>> liklihood) what the user wants. Obviously the user should still have
>>> manual control over this, but if we can "do the right thing"(tm) out of
>>> the box by using this additional metadata then all is well.
>> This would be cool. It seems to me that this metadata is different from
>> the case with Skype. This metadata is more like "preferred sink for a
>> stream." Suppose a game sends audio to 2 gamers. 2 streams, but all have
>> "headset" as their preferred sink.
> Hmmm, not sure what you mean... are there cases where two people play on
> the same machine when they are sharing the same h/w? I appreciate there
> a multi-headed systems but usually the sound h/w is partitioned too and
> thus each user would be logged in separately and run their own instance
> of the software - i.e. two game processes, two separate PA daemon, two
> separate USB headsets etc.
I mean, gamers are playing against each other in 1 game. Buying another 
computer, establishing a network would be a hurdle. I am not a decent 
gamer too. I saw options for 2 players in
http://en.wikipedia.org/wiki/Split_Second:_Velocity
There is a split-screen mode for 2 players with no network. Though I had 
no chance to test it.