This problem showed up while I was trying to get Skype working, but is not Skype specific. Simply put, Kmix recognizes the headset microphone as a Capture device, and the earphones as a Playback device, but nothing I do in Kmix causes sound output to be sent to the earphones. In Skype there is the Make a Test Sound option, and when I activate my system speakers as Playback, I get the expected sound. When I shut them off and activate the headphones, I get nothing. But generally, no sounds come over the head phones when they, and only they, are activated as Playback. For instance, the soundtrack of YouTube videos does not reach the headphones. On the other hand, when I have the headset activated as both Capture and Playback, I can hear myself blowing over the microphone, or tapping on it, suggesting that there is some path to the headphones. Very puzzling. Any suggestions welcomed.
Does it make any difference if you boot with the headphones plugged in? or if you plug them in after booting ?
To look at this in more detail I would need your audio configuration to examine. You can provide that by running the diagnostic script in a terminal with your PC connected to the Internet when USB headphones are plugged in:
/usr/sbin/alsa-info.sh
and select the SHARE/UPLOAD option, and then when the script is complete it will provide in the terminal an internet address/URL to share on this thread, where the audio information has been uploaded to and is now located. That may provide some insight.
I had installed pavucontrol before, but not used it until you suggested it. Using it, and combining yours and JCMcDaniels posts I seem to have solved the problem, by completely disabling every capture and playback device on the system except the headset. This makes Skype work, but is somewhat inconvenient for everyday use when I expect the speakers to work. Part of my problem is that I do not understand the underlying abstract model of the sound system. I assumed that like an old stereo amplifier, there were a number of possible input sources, and then a limited number of output/playback devices, and connecting a particular source to a particular output was a matter of a selection switch for the inputs and either speaker buttons(A or B) or headphone jack, say, for the output. In my mind I had transferred this selection and button paradigm to KMix originally, but that clearly didn’t work this time (although it seems to me it used to). The pavucontrol method seems a bit extreme to me, having to explicitly turn off other components in the Configuration tab.
By the way, this is a desktop machine, OpenSuSE 12.3, 64-bit. The headset is needed because the placement of the speakers and the Logitech integrated camera/microphone means echo-cancellation software works poorly. So I disable the Logitech Webcam Pro 9000 microphone and don’t use the speakers, and obtain very good sound from the headset combination. I am guessing this will not be an issue on my laptop, but I haven’t tried Skype there yet.
Thanks for the help. If there is an SDB article or such, that tells me more about the abstract model, can you point me to it?
Years ago I started an openSUSE sound concept wiki, after which others took over the maintenance / author ship of the wiki (as my paid job is much more busy for me today, than it was 3 to 5 years ago, giving me minimal time to write / maintain wiki). And its my paid job and not my volunteer efforts that pays the rent and puts food on the table.
In addition, as pulse audio was introduced, my knowledge of the openSUSE GNU/Linux sound implementation decreased, and also, the article I started has not been maintained since 2010. You can find it here: SDB:Sound concepts - openSUSE
Likely a newer article is required and I am not aware of one.
Let me say that it makes sense if you have more than one sound device, that you must select the one in which sound is sent to. Else, you need to use one sound device in which you plug your headphones and mic into and assume the built-in speakers are disabled and then earphones/headphones are used instead. PulseAudio could use a feature to send the same audio out to more than one device, but built-in speaker muting may not work. With multiple sound devices, it hard to see how you can satisfy all users wishes with a single setup and so user input may be required as your usage and devices are changed.