Another sound problem (volume and balance)

[BwackNinja]

Sorry in advance, long post ahead! Its a good read though and goes through everything involving Bug 410948.

H'okay, so...

Using Karmic Alpha 4 and have used pulseaudio packages from both the repos and the ubuntu-audio-dev ppa

This problem stems from a discussion on a limitation of alsa:
" * Multichannel mixer controls are not exposed properly: http://mailman.alsa-project.org/pipe...er/011830.html " as seen on http://pulseaudio.org/wiki/AlsaIssues

Discussion took place on the alsa mailing list with:
http://mailman.alsa-project.org/pipe...ay/016992.html
and then the next month:
http://mailman.alsa-project.org/pipe...ne/018710.html

Summed up (because it takes a long time to read it all) Lennart complains that ALSA doesn't expose multichannel mixer controls which makes volume control a pain, hence why it has been half-broken forever unless you've managed to fix yourself a decent setup and asks if it can be fixed. Lennart also inquires about any assumptions that can be made about how to control Master, Front, and PCM channels. A resounding "no" came for valid reasons including it'll break stuff and even if it can be considered the right way then regressions (which only happen when stuff is using ALSA directly or possibly the OSS compatibility layer) are still considered ALSA's fault, which is a bad thing. There's the possibility of adding new functions to the api to maintain compatibility but that's a huge pain to do right as well as choosing how different elements should interact, etc. Also - Master, Front, and PCM channels aren't quite used the same in each driver. Next up, Lennart says that he basically did the whole thing, but in PulseAudio and has the configuration files (the files in /usr/share/pulseaudio/alsa-mixer/paths/ /usr/share/pulseaudio/alsa-mixer/profile-sets/ if you want to try messing with them).

Problems as shown in this report stem from Master being set to be thought of as increasing the volume of all channels (which is true for some but not all) and PCM being thought as increasing the volume of all channels (which is true for some but not all). Changing these to be what you'd consider right for your audio card in /usr/share/pulseaudio/alsa-mixer/paths/analog-output.conf and /usr/share/pulseaudio/alsa-mixer/paths/analog-output.conf.common and moving the channels individually in pavucontrol and watching them change in alsamixer will show that only the channels that you change in pavucontrol will affect the corresponding channels in alsamixer.

Changing these globally will adversely affect hardware that was working properly to begin with, but the framework that Lennart has allows for udev rules to choose a specific configuration file based on the hardware/drivers. A good example is the AC97 cards, which, according to the mailing list, all operate in this same fashion that is broken with the current analog-output.conf and analog-output.conf.common because Master and PCM do not control more than the front. To truly fix this, we would need large scale testing of surround sound setups with different cards and facilitate the correct working of all hardware.

This, while slightly better than the current situation, has one more problem. Currently for me and others subscribed to this bug only one ALSA volume slider is moving at a time so you would simultaneously have output blasting on your rear speakers but nothing coming out of your front speakers. The intended behavior of the new mixer logic in PulseAudio is to move only one at a time and multiply the dB values to achieve a greate level of granularity in volume control without just resorting to doing that fully in hardware. The problem as I see it is that it seems not to be doing this per channel (as you can see when increasing the volume each ALSA volume control will take longer to reach its maximum than the previous one so I'd call this mixer logic mostly-working) but instead is multiplying ALL the dB values of controls that have the "merge" option set.

Finally, there is the problem of whether the dB values are accurate, which is something that Lennart looked at and made a tool for. On my hardware, though the Master and Surround have the same reported dB range (0 to -46.40dB) and PCM has the range 0 to -34.50dB, for the Master (front) to match the loudness of the Surround (rear), I have to set PCM to about 85%, which I don't think makes sense when multiplying the dB values.

I hope this is fixed soon, I like my surround sound

~BwackNinja

[BwackNinja]

Any updates on this? I haven't seen anyone complaining about this anywhere other than the forums and that bug report. Probably because everything is working fine in stereo (which I'm resorting to right now).

[phenest]

Originally Posted by BwackNinja Any updates on this? I haven't seen anyone complaining about this anywhere other than the forums and that bug report. Probably because everything is working fine in stereo (which I'm resorting to right now).

The problem exists in stereo, but it's simply not noticeable. For example, I can plug in my headphones, and everything seems fine. But I can't disable the LFE channel on my laptop, and even if I could, it doesn't solve the problem. The lack of complaints or bug reports simply suggest that the majority of users only use stereo output.

[ma2k1]

Same here
Inspiron 9400, karmic koala alpha 4.
Volume is too loud, if i try to regulate with alsamixer I see that PCM, LFE and MASTER work together... only when PCM and LFE are at 100% Master start to increase...

Any news?