At Mon, 02 Apr 2012 15:39:30 +0200, David Henningsson wrote:
On 03/30/2012 04:31 PM, Takashi Iwai wrote:
At Thu, 29 Mar 2012 22:48:20 +0200, David Henningsson wrote:
The internal mic input is phase inverted on one channel. To avoid people in userspace summing the channels together and get zero result, use a separate mixer control for the inverted channel.
BugLink: https://bugs.launchpad.net/bugs/903853 Signed-off-by: David Henningssondavid.henningsson@canonical.com
Inverted Internal Mic, take 2.
I refined my idea a little. By calling the left channel "Internal Mic" and the right channel "Inverted Internal Mic", we're giving the user a choice whether to activate it or keeping it muted.
For PulseAudio, the quick fix is just to make sure the "Inverted Internal Mic" control is muted. If we in the long run want to actually handle this and swap the phase around, we now also have a reasonable indicator if we should do this, by checking if the "Inverted Internal Mic" control exists.
For the ALSA init DB (i e, should the "Inverted Internal Mic" be initially muted?) that is no big deal for me, since this will be overwritten by PulseAudio later anyway.
I'm also open for better names than "Inverted Internal Mic" if you have any.
As with the previous version, this is a draft version of the patch, and could be done slightly more elegant I guess.
What do you say?
Good, this sounds like a reasonable solution.
I'm happy you like it!
About the name, I'm not sure, too. I'm no good godfather. Maybe other people have a better clue.
Once when the patch is finialized, let me know.
I've been working with it today, and I'm sending a new patch in a separate email.
I think we clean up the fixup functions by moving the alc_fixup_*() as common functions. ALC_FIXUP_SKU can be converted to an ALC_FIXUP_FUNC, then there is no Realtek-specific things there.
I'm happy for all consolidating between codec drivers. Perhaps this could be a good start for a new hda_auto.c file that separates the low-level communication in hda_codec.c, from the auto-parser related stuff which could reside in hda_auto.c?
Yeah, some code reorganization would be needed in near future.
Takashi