January 2010 - Alsa-devel - mailman.alsa-project.org

Re: [alsa-devel] Reg: Audio recording with twl4030 codec
by Peter Ujfalusi 25 Jan '10

25 Jan '10

Hello, Let's put alsa-devel back to the CC... On Monday 25 January 2010 09:32:45 ext shariff wrote: > Hi Peter, > > 1) As we only have one Main MIC when I am recording in stereo mode by > specifying "-c 2", one channel is silent and one channel is recorded > properly. As it should. Just as a sidenote, I'm usually using -f dat for the format, which gives 48KHz, stereo 16LE > > 2) But still the recorded channel is noisy, how can I reduce the noise of > the recorded channel? Hmm, what kind of noise? How are you configuring the dai with the snd_soc_dai_set_fmt? For stereo it should be (if OMAP is slave, TWL is master): SND_SOC_DAIFMT_I2S | SND_SOC_DAIFMT_NB_NF | SND_SOC_DAIFMT_CBM_CFM; For TDM mode (four channel): SND_SOC_DAIFMT_DSP_A | SND_SOC_DAIFMT_IB_NF | SND_SOC_DAIFMT_CBM_CFM; See omap3beagle.c how it is done. > > I have one more question. ( Regarding Audio path) > > ---> If I want to avoid that one channel silent, do I need to connect "SUB > MIC or AUXR"?( as these are connected right ADC). If you want audio data there, than yes you need another mic for the right channel. > ----> If I need to record 4 channel audio without any silent channel do I > need to connect digital microphones also? You can duplicate the channel 1 and 3 to channel 2 and 4, meaning that you route the ADCL and ADCR to these channels (TX1 and TX2 path are both analog). You can also hook up a stereo digital mic to the digimic1 interface, but remember, that the digimic needs 48KHz sampling rate. -- Péter > > Regards, > Shariff > > On Fri, Jan 22, 2010 at 6:29 PM, shariff > <shariffsinbox(a)gmail.com<mailto:shariffsinbox@gmail.com>> wrote: Hi Peter, > > Thank you very much for your suggestions. > > Regards, > Shariff > > > On Fri, Jan 22, 2010 at 6:11 PM, Peter Ujfalusi > <peter.ujfalusi(a)nokia.com<mailto:peter.ujfalusi@nokia.com>> wrote: Hello, > > On Friday 22 January 2010 11:57:23 ext shariff wrote: > > Hi all, > > > > I am working with OMAP3 based custom board that has twl4030 codec > > & using linux-2.6.29 kernel from opensource. > > This is kind of old... > > > I have some queries regarding Audio recording. > > > > Audio codec is connected to twl4030 as bellow > > > > TWL4030 > > > > |--------------------| > > | > > | |------ MAINMIC > > > > McBSP2 ------ |i2s/tdm | > > > > | ___________| > > > > --> One Analog MIC is connected to Main Mic and I am using Audio > > path not the Voice path. > > > > After selecting Main Mic through amixer control. > > > > and recording audio using "arecord -c 1 -r 8000 -f S16_LE -t > > wav -d 20 record-file.wav" > > The hardware should have been opened in stereo mode, since it is not > supporting the mono, you can verify it by appending -v to arecord and see > how the HW has been opened at the end. > > > --> Here record is successful but with lot of noise. > > > > 1) Can we record Mono(1 channel) through I2S from TWL4030 > > codec? > > The TWL codec "Audio" path is either stereo or 4 channel. > You need to record stereo/4 channel, and than extract/use the data from the > valid channel. > In case of Main mic it is the left channel, or Channel 1 or/and 2 > (depending on the routings) in four channel mode. > > > 2) To record stereo(2 channels), do we need two Mic's ? (i.e do > > I need Sub Mic also). > > No you don't need. The right channel will not contain valid data (most > likely it is going to be silent). > > You need to enable the Mic bias 1 for the main mic in your machine driver, > if you have not done that yet: > > static const struct snd_soc_dapm_widget your_widgets[] = { > ... > SND_SOC_DAPM_MIC("Your Mic", NULL), > ... > }; > > static const struct snd_soc_dapm_route your_audio_map[] = { > ... > {"MAINMIC", NULL, "Mic Bias 1"}, > {"Mic Bias 1", NULL, "Your Mic"}, > ... > }; > > I'm not sure about the state of the twl soc codec driver in 2.6.29, but my > guess is that it should work. > > Are there any capture routing option for TX1 path, which selects between > Analog and Digimic0? If there is than that has to be set to analog as > well. > > > Regards, > > Shariff > > -- > Péter >

2 1

[alsa-devel] ASoC updates for 2.6.34
by Mark Brown 25 Jan '10

25 Jan '10

The following changes since your topic/asoc branch are available in the git repository at: git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound-2.6.git for-2.6.34 Since they include (via the 2.6.33 merge) a merge of Linus' tree I've only included the changes that aren't already in your tree or mainline in the summary below, though it looks like the diffstat picked up some extra changes due to the way I tried to do that. Barry Song (4): ASoC: ad1836: reset and restore clock control mode in suspend/resume entry ASoC: ad1938: fix typo, rename mask to rx_mask for ad1938_set_tdm_slot ASoC: ad1938: let soc-core dapm handle PLL power ASoC: soc-cache: cleanup training whitespace and coding style Ilkka Koskinen (2): ASoc: tpa6130a2: Remove unnecessary variable ASoC: tpa6130a2: Support for tpa6140's regulators Kuninori Morimoto (3): ASoC: fsi: Add over_period flag to prevent the misunderstanding ASoC: fsi: Add fsi_get_dai to get snd_soc_dai ASoC: fsi: Add over/under run error settlement Mark Brown (9): ASoC: Remove unneeded suspend checks from CODEC drivers ASoC: Use snprintf() when generating stream names ASoC: Fix variable shadowing warning in TLV320AIC3x ASoC: Only restore non-default registers for WM8731 ASoC: Only restore non-default registers for WM8776 ASoC: Only restore non-default registers for WM8961 ASoC: Implement suspend and resume for WM8993 ASoC: Fix WM8350 DSP mode B configuration Merge branch 'for-2.6.33' into for-2.6.34 Peter Ujfalusi (5): ASoC: tlv320dac33: Change nsample switch to FIFO mode enum ASoC: tlv320dac33: Introduce prefill and playback state handlers ASoC: tlv320dac33: Clean up the hardware configuration code ASoC: tlv320dac33: Add new FIFO mode: mode 7 ASoC: tlv320dac33: Safety check for codec slave mode include/sound/soc.h | 3 + include/sound/tpa6130a2-plat.h | 6 + include/sound/wm8904.h | 57 + include/sound/wm8955.h | 26 + sound/soc/codecs/Kconfig | 12 + sound/soc/codecs/Makefile | 10 +- sound/soc/codecs/ad1836.c | 32 + sound/soc/codecs/ad1938.c | 64 +- sound/soc/codecs/cs4270.c | 43 +- sound/soc/codecs/da7210.c | 589 +++++++++ sound/soc/codecs/da7210.h | 24 + sound/soc/codecs/tlv320aic3x.c | 75 +- sound/soc/codecs/tlv320dac33.c | 282 ++++- sound/soc/codecs/tpa6130a2.c | 109 ++- sound/soc/codecs/wm8350.c | 2 +- sound/soc/codecs/wm8727.c | 66 +- sound/soc/codecs/wm8731.c | 3 + sound/soc/codecs/wm8753.c | 8 - sound/soc/codecs/wm8776.c | 2 + sound/soc/codecs/wm8904.c | 2556 +++++++++++++++++++++++++++++++++++++ sound/soc/codecs/wm8904.h | 1681 ++++++++++++++++++++++++ sound/soc/codecs/wm8955.c | 1151 +++++++++++++++++ sound/soc/codecs/wm8955.h | 489 +++++++ sound/soc/codecs/wm8961.c | 3 + sound/soc/codecs/wm8990.c | 8 - sound/soc/codecs/wm8993.c | 67 + sound/soc/davinci/davinci-mcasp.c | 18 +- sound/soc/davinci/davinci-mcasp.h | 1 + sound/soc/davinci/davinci-pcm.c | 2 +- sound/soc/s3c24xx/s3c64xx-i2s.c | 120 +-- sound/soc/sh/Kconfig | 8 + sound/soc/sh/Makefile | 2 + sound/soc/sh/fsi-da7210.c | 83 ++ sound/soc/sh/fsi.c | 164 ++- sound/soc/soc-cache.c | 2 +- sound/soc/soc-core.c | 11 +- sound/soc/soc-dapm.c | 37 +- 37 files changed, 7457 insertions(+), 359 deletions(-) create mode 100644 include/sound/wm8904.h create mode 100644 include/sound/wm8955.h create mode 100644 sound/soc/codecs/da7210.c create mode 100644 sound/soc/codecs/da7210.h create mode 100644 sound/soc/codecs/wm8904.c create mode 100644 sound/soc/codecs/wm8904.h create mode 100644 sound/soc/codecs/wm8955.c create mode 100644 sound/soc/codecs/wm8955.h create mode 100644 sound/soc/sh/fsi-da7210.c

4 19

[alsa-devel] ALSA: hda - add possibility to choose speakers configuration for 4930g
by Łukasz Wojniłowicz 25 Jan '10

25 Jan '10

I've got such possibility on MS Windows XP with Realtek drivers (prove is at http://img694.imageshack.us/i/choosechannelmod.png/ ) so I should have possibility to do it also on Linux e.g. here http://img704.imageshack.us/i/screenshotsoundpreferen.png/ Patch given in attachment should do the thing. patch_realtek.c was downloaded from http://git.alsa-project.org/?p=alsa-kernel.git on Thu, 21 Jan 2010 13:58:25 +0000

2 1

Re: [alsa-devel] snd_pcm_hw_params_free() immediately after snd_pcm_hw_params() ???
by Clemens Ladisch 24 Jan '10

24 Jan '10

Raymond Yau wrote: > Is there any flags in substream which can indicate snd_xxx_hw_params() was > called at first time since snd_xxx_hw_params() were called many times > during OSS emulation (probing the supported rates , channels , .... ) ? There is no such flag. It is the responsibility of the driver to track the resources it has allocated. If a resource does not depend on the hardware parameters, it might be a better idea to allocate/free it in the open/close callbacks. Best regards, Clemens

3 2

[alsa-devel] [PATCH] add simplified versions of widget macros
by Guennadi Liakhovetski 23 Jan '10

23 Jan '10

Many macros from include/sound/soc-dapm.h take an array and a number of elements in it as arguments, whereas most users use static arrays and use "x, ARRAY_SIZE(x)" as arguments. This patch adds simplified versions of those macros, calling ARRAY_SIZE() internally. Signed-off-by: Guennadi Liakhovetski <g.liakhovetski(a)gmx.de> --- diff --git a/include/sound/soc-dapm.h b/include/sound/soc-dapm.h index c5c95e1..c0922a0 100644 --- a/include/sound/soc-dapm.h +++ b/include/sound/soc-dapm.h @@ -95,6 +95,21 @@ .shift = wshift, .invert = winvert, .kcontrols = wcontrols, \ .num_kcontrols = 1} +/* Simplified versions of above macros, assuming wncontrols = ARRAY_SIZE(wcontrols) */ +#define SOC_PGA_ARRAY(wname, wreg, wshift, winvert,\ + wcontrols) \ +{ .id = snd_soc_dapm_pga, .name = wname, .reg = wreg, .shift = wshift, \ + .invert = winvert, .kcontrols = wcontrols, .num_kcontrols = ARRAY_SIZE(wcontrols)} +#define SOC_MIXER_ARRAY(wname, wreg, wshift, winvert, \ + wcontrols)\ +{ .id = snd_soc_dapm_mixer, .name = wname, .reg = wreg, .shift = wshift, \ + .invert = winvert, .kcontrols = wcontrols, .num_kcontrols = ARRAY_SIZE(wcontrols)} +#define SOC_MIXER_NAMED_CTL_ARRAY(wname, wreg, wshift, winvert, \ + wcontrols)\ +{ .id = snd_soc_dapm_mixer_named_ctl, .name = wname, .reg = wreg, \ + .shift = wshift, .invert = winvert, .kcontrols = wcontrols, \ + .num_kcontrols = ARRAY_SIZE(wcontrols)} + /* path domain with event - event handler must return 0 for success */ #define SND_SOC_DAPM_PGA_E(wname, wreg, wshift, winvert, wcontrols, \ wncontrols, wevent, wflags) \ @@ -126,6 +141,23 @@ .invert = winvert, .kcontrols = wcontrols, .num_kcontrols = 1, \ .event = wevent, .event_flags = wflags} +/* Simplified versions of above macros, assuming wncontrols = ARRAY_SIZE(wcontrols) */ +#define SOC_PGA_E_ARRAY(wname, wreg, wshift, winvert, wcontrols, \ + wevent, wflags) \ +{ .id = snd_soc_dapm_pga, .name = wname, .reg = wreg, .shift = wshift, \ + .invert = winvert, .kcontrols = wcontrols, .num_kcontrols = ARRAY_SIZE(wcontrols), \ + .event = wevent, .event_flags = wflags} +#define SOC_MIXER_E_ARRAY(wname, wreg, wshift, winvert, wcontrols, \ + wevent, wflags) \ +{ .id = snd_soc_dapm_mixer, .name = wname, .reg = wreg, .shift = wshift, \ + .invert = winvert, .kcontrols = wcontrols, .num_kcontrols = ARRAY_SIZE(wcontrols), \ + .event = wevent, .event_flags = wflags} +#define SOC_MIXER_NAMED_CTL_E_ARRAY(wname, wreg, wshift, winvert, \ + wcontrols, wevent, wflags) \ +{ .id = snd_soc_dapm_mixer, .name = wname, .reg = wreg, .shift = wshift, \ + .invert = winvert, .kcontrols = wcontrols, \ + .num_kcontrols = ARRAY_SIZE(wcontrols), .event = wevent, .event_flags = wflags} + /* events that are pre and post DAPM */ #define SND_SOC_DAPM_PRE(wname, wevent) \ { .id = snd_soc_dapm_pre, .name = wname, .kcontrols = NULL, \

3 2

[alsa-devel] [PATCH] ASoC: fix a memory-leak in wm8903
by Guennadi Liakhovetski 23 Jan '10

23 Jan '10

Remember to free the temporary register-cache. Signed-off-by: Guennadi Liakhovetski <g.liakhovetski(a)gmx.de> --- should go in -rc, maybe also for stable diff --git a/sound/soc/codecs/wm8903.c b/sound/soc/codecs/wm8903.c index ce5515e..3595bd5 100644 --- a/sound/soc/codecs/wm8903.c +++ b/sound/soc/codecs/wm8903.c @@ -1504,7 +1504,7 @@ static int wm8903_resume(struct platform_device *pdev) struct i2c_client *i2c = codec->control_data; int i; u16 *reg_cache = codec->reg_cache; - u16 *tmp_cache = kmemdup(codec->reg_cache, sizeof(wm8903_reg_defaults), + u16 *tmp_cache = kmemdup(reg_cache, sizeof(wm8903_reg_defaults), GFP_KERNEL); /* Bring the codec back up to standby first to minimise pop/clicks */ @@ -1516,6 +1516,7 @@ static int wm8903_resume(struct platform_device *pdev) for (i = 2; i < ARRAY_SIZE(wm8903_reg_defaults); i++) if (tmp_cache[i] != reg_cache[i]) snd_soc_write(codec, i, tmp_cache[i]); + kfree(tmp_cache); } else { dev_err(&i2c->dev, "Failed to allocate temporary cache\n"); }

3 2

[alsa-devel] ca0106 broken (fwd)
by Russell Whitaker 23 Jan '10

23 Jan '10

Hi Sent msg below and no responce. As a work-around replaced ng code with know good code into 2.6.29.6 and it worked. Tried same trick with 2.6.31.12 and it won't compile. If you send me something to try will do it and report back. Russ ---------- Forwarded message ---------- Date: Wed, 21 Oct 2009 14:54:45 -0700 (PDT) From: Russell Whitaker <russ(a)ashlandhome.net> To: James(a)superbug.demon.co.uk Subject: ca0106 broken Hi ca0106 is known to work in these releases: linux-2.6.27.20 linux-2.6.27.23 linux-2.6.28.10 It is broken in: linux-2.6.29.6 linux-2.6.30.9 For the broken releases dmesg reports: unknown hardware "CA0106" "" "" "" "" Anything I can do to help? Russ

2 3

Re: [alsa-devel] Echo Indigo card on Sabayon Gentoo
by S. Aguinaga 23 Jan '10

23 Jan '10

Hello Giuliano, No luck, same behavior. I edited the first file listed listed below and recompiled it and installed it. Then I restarted the system with the indigo card in place, and the VU meters show levels in the -120 to -100 dB? I do have alsa-driver-1.0.22 my kernel is 2.6.31 // Salvador ________________________________ From: Giuliano Pochini <pochini(a)shiny.it> To: S. Aguinaga <sa_jr(a)yahoo.com> Sent: Sun, January 17, 2010 4:04:38 PM Subject: Re: [alsa-devel] Echo Indigo card on Sabayon Gentoo On Thu, 14 Jan 2010 12:13:55 -0800 (PST) "S. Aguinaga" <sa_jr(a)yahoo.com> wrote: > Hello Fellows, > Just downloaded the latest (released) alsa-lib, alsa-driver, and alsa-utils, but my Indigo card does not work. It has been tested on the same system using the same drivers and on the same system using Fedora 10, that works fine. First install a recent version of echomixer. If you are using alsa-driver 1.0.22 or kernel 2.6.32, try to add this line: #define ECHOCARD_HAS_LINE_OUT_GAIN in alsa-driver-1.0.22/alsa-kernel/pci/echoaudio/indigoio.c or in linux-2.6.32/sound/pci/echoaudio/indigoio.c and recompile. It re-enables the line-out control that I removed last year because it was not supposed to exist in that card. -- Giuliano.

2 2

[alsa-devel] [PATCH] modem.conf Off-hook improve behavior
by David Fries 23 Jan '10

23 Jan '10

From: David Fries <david(a)fries.net> Only restore the old value if it differs from the requested value, because if it has changed restoring the old value overrides the change. Take for example, a voice modem with a .conf that sets preserve off-hook. Start playback (on-hook to off-hook), start record (off-hook to off-hook), stop playback (off-hook to restore on-hook), stop record (on-hook to restore off-hook), Clearly you don't want to leave the modem "on the phone" now that there isn't any playback or recording active. Signed-off-by: David Fries <david(a)fries.net> --- Comments welcome. The default modem.conf has, hooks.0 { type ctl_elems hook_args [ { name "Off-hook Switch" preserve true value "on" lock false optional true } ] } I wrote the patch to be, if you didn't modify it, don't restore the value. I also thought that it might work to check to see if the current mixer value matches the requested value (which is the value the program set on open), and only set the old value if it hasn't been externally modified. The theory being if something else changed the mixer value, we don't want to by default overwrite that change. Some additional flags could also be added to leave the current behavior alone and add the if you didn't modify it, don't restore it, or if it changed after you set it, don't restore it. Here are some flag names if that would be preferred, preserve_if_different restore_if_unmodified no_squash restore_overrides preserve_if_changed. Setting the lock isn't really an option, because then you can't manually "hang up" the modem if you needed to. On a side note, do any of the alsa voice/software modems support notifying userspace when the line is ringing? I'm using intel8x0m, not that I think it matters here. diff --git a/include/control.h b/include/control.h index 2361dc3..ded884a 100644 --- a/include/control.h +++ b/include/control.h @@ -419,6 +419,7 @@ int snd_ctl_elem_value_malloc(snd_ctl_elem_value_t **ptr); void snd_ctl_elem_value_free(snd_ctl_elem_value_t *obj); void snd_ctl_elem_value_clear(snd_ctl_elem_value_t *obj); void snd_ctl_elem_value_copy(snd_ctl_elem_value_t *dst, const snd_ctl_elem_value_t *src); +int snd_ctl_elem_value_compare(snd_ctl_elem_value_t *left, const snd_ctl_elem_value_t *right); void snd_ctl_elem_value_get_id(const snd_ctl_elem_value_t *obj, snd_ctl_elem_id_t *ptr); unsigned int snd_ctl_elem_value_get_numid(const snd_ctl_elem_value_t *obj); snd_ctl_elem_iface_t snd_ctl_elem_value_get_interface(const snd_ctl_elem_value_t *obj); diff --git a/src/control/control.c b/src/control/control.c index df249dd..159b09a 100644 --- a/src/control/control.c +++ b/src/control/control.c @@ -2256,6 +2256,18 @@ void snd_ctl_elem_value_copy(snd_ctl_elem_value_t *dst, const snd_ctl_elem_value } /** + * \brief compare one #snd_ctl_elem_value_t to another + * \param dst pointer to destination + * \param src pointer to source + * \return 0 on match, less than or greater than otherwise, see memcmp + */ +int snd_ctl_elem_value_compare(snd_ctl_elem_value_t *left, const snd_ctl_elem_value_t *right) +{ + assert(left && right); + return memcmp(left, right, sizeof(*left)); +} + +/** * \brief Get CTL element identifier of a CTL element id/value * \param obj CTL element id/value * \param ptr Pointer to returned CTL element identifier diff --git a/src/control/setup.c b/src/control/setup.c index 408244e..3606daf 100644 --- a/src/control/setup.c +++ b/src/control/setup.c @@ -192,7 +192,17 @@ int snd_sctl_remove(snd_sctl_t *h) return err; } } - if (elem->preserve) { + /* Only restore the old value if it differs from the requested + * value, because if it has changed restoring the old value + * overrides the change. Take for example, a voice modem with + * a .conf that sets preserve off-hook. Start playback (on-hook + * to off-hook), start record (off-hook to off-hook), stop + * playback (off-hook to restore on-hook), stop record (on-hook + * to restore off-hook), Clearly you don't want to leave the + * modem "on the phone" now that there isn't any playback or + * recording active. + */ + if (elem->preserve && snd_ctl_elem_value_compare(elem->val, elem->old)) { err = snd_ctl_elem_write(h->ctl, elem->old); if (err < 0) { SNDERR("Cannot restore ctl elem");

2 2

[alsa-devel] default alsa pcm improvement
by Aleksandar Kostadinov 23 Jan '10

23 Jan '10

Hello, _From some time now ALSA is a real pleasure to use. Namely it works by OOB and has lots of predefined useful PCMs so one doesn't need to be a guru to do advanced operations. I think that a small improvement in how default PCM is defined can help users with multiple sound cards that like to use pure ALSA and not some of the many available sound servers. I want to be able to switch the default sound device easily in asoundrc but the way the default pcm is defined makes this not very straightforward. I like to use the definitions of default:<CARD> but if I define "pcm.!default" in .asoundrc I lose the "default:<CARD>" definition. I now can rename the "default" definition in default.conf to someting like "defaultorg" and then reference it asoundrc but that requires root as well change can possibly disappear after an alsa-lib package upgrade. Or I can use the dmix and dsnoop predefined PCMs to put together PCMs that I can later reference in my "pcm.!default" definition but in this case one needs to know if the card needs dsnoop/dmix and probably other tweaks that default.conf contains. Additionally this needs to be done for every sound card one owns. So I would suggest the "default" PCM as defined now in the default.conf file to be renamed to something like "alsadefault". Then "default" PCM should be an alias to the alsadefault pcm. This will be backwards compatible to what users have now and owners of multiple sound devices can define their "default" pcm in asoundrc referencing the alsadefault pcm benefiting from the good work already done and not have to mess with a complex asoundrc file or make changes to the system configurations files. Please CC me in replies because I'm not subscribed. Thanks again for the cool work!

2 2