Comment 19 for bug 1236965

After composing a long and thoughful comment, with incredible pearls of wisdom, including a short and easy-to-implement plan for sustainable peace in the middle east within the year, I clicked the button to attach a screenshot, and got an error HTTP 414 Request-URI Too Large, The requested URL's length exceeds the capacity limit for this server. Apache/2.2.14 (Ubuntu) Server at api.launchpad.net Port 443. When I clicked the back-button, the page reloaded, and -- although firefox *tries* to retain textarea content for me -- launchpad managed to delete it. I was so angry, that I have now completely forgotten the inspired details of my peace plan! Please fix this thing, so that next time inspiration strikes me, your software will not be responsible for endless turmoil and bloodshed and bitterness. Workaround: always remember to click attach-button *first* and then and only then compose long thoughtful comment, which works fine. While I'm complaining: why is there no button to include multiple attachments in a single submission? Makes me have to piddle around zipping up my screenshots (or pasting multiple screenshots into a bigger PNG that is hand-edited...) or submit each upload as a separate comment. I also want a pony. https://en.wikipedia.org/wiki/WP:PONY

Since I composed some of my commentary in an external text-editor, here you go. I'm still working my way through your questions; these are just some quick answers to the low-hanging fruit. Gratitude for your help.

"David: worried about that you can't seem to get sound out of anything but the front speakers, is this correct?"
No... not really. I can get sound out of 'Center' as well, in some tests. I also get considerably-quieter-but-still-audible sound out of 'side&rear_left&right' named channels... but note that only five+plus+subwoofer physical speakers exist... not sure how to interpret that mismatch. It is difficult for me to pinpoint the physical location of 'side&rear_left&right' audio, but it is not coming from frontLeft or frontRight, that much is clear. ('Center' also seems distinct.) That said, I'm pretty sure I never get any audio at all from the subwoofer, in tests I have run so far. I removed the -v argswitch from your test; not supported on my box.

$ speaker-test --help
speaker-test 1.0.25
Usage: speaker-test [OPTION]...
-h,--help help
-D,--device playback device
-r,--rate stream rate in Hz
-c,--channels count of channels in stream
-f,--frequency sine wave frequency in Hz
-F,--format sample format
-b,--buffer ring buffer size in us
-p,--period period size in us
-P,--nperiods number of periods
-t,--test pink=use pink noise, sine=use sine wave, wav=WAV file
-l,--nloops specify number of loops to test, 0 = infinite
-s,--speaker single speaker test. Values 1=Left, 2=right, etc
-w,--wavfile Use the given WAV file as a test sound
-W,--wavdir Specify the directory containing WAV files
Recognized sample formats are: S8 S16_LE S16_BE FLOAT_LE S32_LE S32_BE

$ speaker-test -D plughw:PCH -t wav -c 6
speaker-test 1.0.25
Playback device is plughw:PCH
Stream parameters are 48000Hz, S16_LE, 6 channels
WAV file(s)
Rate set to 48000Hz (requested 48000Hz)
Buffer size range from 64 to 16384
Period size range from 32 to 8192
Using max buffer size 16384
Periods = 4
was set period_size = 4096
was set buffer_size = 16384
 0 - Front Left [loud, from frontLeft located southwest of Z-key on keyb.]
 4 - Center [nothing]
 1 - Front Right [loud, from frontRight located southeast of M-key on keyb.]
 3 - Rear Right [nothing]
 2 - Rear Left [nothing]
 5 - LFE [nothing]
Time per period = 8.363885^C

Note that e.g. my 9ch test-run *does* produce sound from center/rearRight/rearLeft speakers.

"Raymond: use hda-jack-retask to remove the line out jack and force the driver to put three speakers in multi_outs"
I have heard of hda-jack-retask , and it may help us here. What exactly did you have in mind? Do you want me to run some specific command, and then run some specific test?

"Raymond: retask line out is not sufficicent, must also use hint to disable"
Do you want me to put the text you posted into some file somewheres? (multi_io= false \n autoconfig: ... Best config: ...)

"Raymond: the driver need to assign DAC to 5.1/7.1 output before assign DAC to the headphone"
So, should I patch some file, to remove the snd_kcontrol_new alc883_targa_mixer lines of code? Steps please?

"Takashi in Feb'12: ...auto-parser now. We can remove all alc*_quirks.c"
- targa-8ch-dig Targa/MSI with 8-channel (MSI GX620)
- laptop-eapd 3-jack with SPDIF I/O and EAPD (Clevo M540JE, M550JE)
- clevo-m540r Clevo M540R (6ch + digital)
- clevo-m720 Clevo M720 laptop series

What is this auto-parser that Takashi used to eliminate quirks for targa-8ch-dig and the various Clevo ALC883 units? Does my audio-hardware fail to work with the auto-parser, and thus I need some kind of driver-quirk? I am new to this audio-debugging world, as you can probably tell. If this question has no simple answer, can you point me to the relevant helpdocs where I should start, so as to eventually be able to answer it myself, once I grok how everything works a bit better?

https://git.kernel.org/cgit/linux/kernel/git/tiwai/sound.git/tree/Documentation/sound/alsa/HD-Audio-Models.txt
ALC88x/898/1150 [presumably these are ALC883 units not handled by the auto-parser?]
======================
  acer-aspire-4930g Acer Aspire 4930G/5930G/6530G/6930G/7730G
  acer-aspire-8930g Acer Aspire 8330G/6935G
  acer-aspire Acer Aspire others
  inv-dmic Inverted internal mic workaround
  no-primary-hp VAIO Z/VGC-LN51JGB workaround (for fixed speaker DAC)
ALC66x/67x/892 [here are some ALC892 units not handled by the auto-parser?]
==============
  mario Chromebook mario model fixup
  asus-mode1 ASUS
  asus-mode2 ASUS
  asus-mode3 ASUS
  asus-mode4 ASUS
  asus-mode5 ASUS
  asus-mode6 ASUS
  asus-mode7 ASUS
  asus-mode8 ASUS
  inv-dmic Inverted internal mic workaround
  dell-headset-multi Headset jack, which can also be used as mic-in

Am I supposed to have my model listed in the ALC892 section, at some point? Or are we thinking the auto-parser handles my hardware, and the difficulty lies elsewhere, e.g. in the dotconf changes David wants me to try? Or some combo?

Note that, at the moment, I personally am not attempting to accomplish 7.1 output through the 3.5mm quad-jacks. I'm not sure I have the hardware to test that, actually... do I just need some external "cheap pc speaker" type stuff, or do I need something special (like a fancy audio-system) to plug into the 3.5mm-spdif-out-jack? Maybe there is a way to test that is works, i.e. some software that can monitor the volume-levels going to the 7.1 config, without actually plugging in appropriate hardware... what might the name of this software be, if it exists?

Personally, I'll be happy to get the onboard 5.1 speakers working properly. Originally, they were not recognized by pulseaudio at all, and did not show up in sound-settings. After running the workaround in comment#32 on the other thread, now the onboard speakers do show up. However, using the regular volume-control is a big fail: pulseaudio is only moving the alsamixer 'main' and alsamixer 'pcm' channels up and down, ignoring all the others. See attached PNG (sorry the plan for world peace isn't found therein). At present, most of the power of the onboard speakers is either wasted at high aggregate volumes (if I use alsamixer to set the ignored-by-pulseaudio channles manually to a low volume), or overpowering at low aggregate volumes (if I instead use alsamixer to set the ignored-by-pulseaudio channels manually to a high volume). That said, I'm happy to help, and see if we can make 7.1 analog out function properly, so that other people with similar hardware can enjoy it, but for me personally the onboard 5.1 working properly is my primary goal.