[pulseaudio-discuss] Forcing a mapping from a stream to a device

Colin Guthrie gmane at colin.guthr.ie
Wed Dec 15 02:56:47 PST 2010


[Adding the list back in - I think I may have removed it accidentally :s ]

'Twas brillig, and Matt Feifarek at 15/12/10 00:06 did gyre and gimble:
> On Thu, Dec 9, 2010 at 12:32 PM, Colin Guthrie <gmane at colin.guthr.ie
>     I've a funny feeling we discussed this on IRC but perhaps not.
> 
> 
> Jeez, I hope not. I did have this problem before with another device...
> I may have asked... if so, I didn't solve the problem, and therefore
> forgot. If I did, I really am sorry to waste your time.


Oh don't worry. I just remember discussing something similar with
someone recently and with the length of time this mail was stuck in the
moderation queue, I figured it may have been you. No need to apologise
for wasting time etc. I wouldn't have replied if I didn't want to :D


>     Can you provide output "pacmd list-sink-inputs" output when one of these
>     calls is in progress. The interesting bit will be the stream restore id
>     in the proplist.
> 
> 
> Attached. It seems that Chromium might be shooting straight for ALSA
> rather than Pulse. This ought to be caught by pulse, right?

If it's listed at all in pacmd it definitely is "caught by pulse". The
Chromium connection going via alsa is perfectly OK. It's basically just
using alsa API and using the 'default' device which is correctly routed
via pulse. Technically there are several different parts to "alsa" (i.e
the driver and the audio library) Using Alsa the api to talk to pulse is
a proper part of a good setup :)


>     This will be more complicated as the browser is a general purpose
>     system, and yet it's trying to do use-case specific things.
> 
> 
> Well, perhaps in my case I'm lucky... it's a specific plugin, that seems
> to be properly identifying itself: "GoogleTalkPlugin".

Yeah that's good. It will be saved/restored independently of the flash
and movies and such like coming from Chromium.

It doesn't set a media.role for itself which it likely should.

>     Skype should work fine. Skype even identifies it's streams properly -
>     "phone" for it call audio and "event" for the ringing sound etc.
> 
> 
> Hmm. I'm not having good luck, but I'll keep trying. Good to know.

Have a look at pacmd list-sink-inputs (and list-source-outputs) when a
Skype call is in progress. You should see a "media.role = phone"  in
there somehow.

>     Generally speaking, pavucontrol is the app you want. I have no idea what
>     earcandy is.
> 
> 
> It's not really worth mucking with, but:
> https://launchpad.net/earcandy

/me has heard of it actually... I just keep forgetting as it's not on my
radar 99% of the time. I like the name tho' :D


> Glad to hear that the stock conf should be working for me; now at least
> I know it's worth it to keep trying.
> 
> Thanks so much for your detailed response.

No worries, feel free to ask more questions if things don't behave
better for you. Generally speaking "pacmd list" output is the most
useful debug - preferably when the things that's not behaving is active
in some way.

Cheers

Col



-- 

Colin Guthrie
gmane(at)colin.guthr.ie
http://colin.guthr.ie/

Day Job:
  Tribalogic Limited [http://www.tribalogic.net/]
Open Source:
  Mageia Contributor [http://www.mageia.org/]
  PulseAudio Hacker [http://www.pulseaudio.org/]
  Trac Hacker [http://trac.edgewall.org/]



More information about the pulseaudio-discuss mailing list