Comment 156 for bug 116326

Ben,

I am willing to help to test this. I have a similar laptop to the one of venky (toshiba U305),
vendor-id 0x1179, 0xff50, and I am running Feisty with a vanilla (Linus') 2.6.23-rc9 kernel, plus I have tried various alsa drivers.
I want to be sure to do the right thing, so...

 * I recompile the drivers (./configure --with-cards=hda-intel)

 * I have alsa lib 1.0.14 installed; someone on the lkml list told me that's ok to test reinstalling just the kernel modules, is it true?

 * My sound sort-of-work with 1.0.15-rcX, last hg snapshot (20071002), and the last realtek driver (4.06c), autodetected as model=toshiba. The problem is with the very lousy quality of recording, given in my opinion for the persistent problem of the huge dc offset in input (there are snaphsot and plenty of details on the alsa-bug 3104), which is not resolved.

 * I tried to test model=3stack and model=acer with the procedure:
    - Install the new driver (as above)
    - /etc/init.d alsasound stop (a couple of times, first one fails)
    - modify /etc/modprobe.d/alsa-base adding model=... to the options snd-hda-intel line
    - restart alsa by modprobe snd-hda-intel (and resetting and setting all the mixer values).

Till now, no drivers has the sufficient quality in microphone input to make a Skype call possible.
If I can help test patches, please tell me. Should I try 5stack? Is the aforemention procedure correct for testing?

Romano