[alsa-devel] RFC: support for 12 & 24Khz
Hey Takashi,
For compressed audio we also need to support the PCM rates of 12 and 24KHz.
Looking at pcm.h these are not defined so we can simply add them at the end. But am worried about wider impact of adding these rates.
Can you let me know if more is required to be done or below is fine
diff --git a/include/sound/pcm.h b/include/sound/pcm.h index 84b10f9..e418d8d 100644 --- a/include/sound/pcm.h +++ b/include/sound/pcm.h @@ -126,6 +126,8 @@ struct snd_pcm_ops { #define SNDRV_PCM_RATE_96000 (1<<10) /* 96000Hz */ #define SNDRV_PCM_RATE_176400 (1<<11) /* 176400Hz */ #define SNDRV_PCM_RATE_192000 (1<<12) /* 192000Hz */ +#define SNDRV_PCM_RATE_12000 <1<<13> /* 12000Hz */ +#define SNDRV_PCM_RATE_24000 <1<<14> /* 24000Hz */
#define SNDRV_PCM_RATE_CONTINUOUS (1<<30) /* continuous range */ #define SNDRV_PCM_RATE_KNOT (1<<31) /* supports more non-continuos rates */ diff --git a/sound/core/pcm_native.c b/sound/core/pcm_native.c index a68d4c6..42600b0 100644 --- a/sound/core/pcm_native.c +++ b/sound/core/pcm_native.c @@ -1779,12 +1779,13 @@ static int snd_pcm_hw_rule_sample_bits(struct snd_pcm_hw_params *params, return snd_interval_refine(hw_param_interval(params, rule->var), &t); }
-#if SNDRV_PCM_RATE_5512 != 1 << 0 || SNDRV_PCM_RATE_192000 != 1 << 12 +#if SNDRV_PCM_RATE_5512 != 1 << 0 || SNDRV_PCM_RATE_24000 != 1 << 14 #error "Change this table" #endif
static unsigned int rates[] = { 5512, 8000, 11025, 16000, 22050, 32000, 44100, - 48000, 64000, 88200, 96000, 176400, 192000 }; + 48000, 64000, 88200, 96000, 176400, 192000, + 12000, 24000 };
const struct snd_pcm_hw_constraint_list snd_pcm_known_rates = { .count = ARRAY_SIZE(rates),
~Vinod --
On 07/24/2013 07:26 PM, Vinod Koul wrote:
Hey Takashi,
For compressed audio we also need to support the PCM rates of 12 and 24KHz.
Looking at pcm.h these are not defined so we can simply add them at the end. But am worried about wider impact of adding these rates.
Can you let me know if more is required to be done or below is fine
diff --git a/include/sound/pcm.h b/include/sound/pcm.h index 84b10f9..e418d8d 100644 --- a/include/sound/pcm.h +++ b/include/sound/pcm.h @@ -126,6 +126,8 @@ struct snd_pcm_ops { #define SNDRV_PCM_RATE_96000 (1<<10) /* 96000Hz */ #define SNDRV_PCM_RATE_176400 (1<<11) /* 176400Hz */ #define SNDRV_PCM_RATE_192000 (1<<12) /* 192000Hz */ +#define SNDRV_PCM_RATE_12000 <1<<13> /* 12000Hz */ +#define SNDRV_PCM_RATE_24000 <1<<14> /* 24000Hz */
You probably meant (1<<13), not <1<<13> ?
#define SNDRV_PCM_RATE_CONTINUOUS (1<<30) /* continuous range */ #define SNDRV_PCM_RATE_KNOT (1<<31) /* supports more non-continuos rates */ diff --git a/sound/core/pcm_native.c b/sound/core/pcm_native.c index a68d4c6..42600b0 100644 --- a/sound/core/pcm_native.c +++ b/sound/core/pcm_native.c @@ -1779,12 +1779,13 @@ static int snd_pcm_hw_rule_sample_bits(struct snd_pcm_hw_params *params, return snd_interval_refine(hw_param_interval(params, rule->var), &t); }
-#if SNDRV_PCM_RATE_5512 != 1 << 0 || SNDRV_PCM_RATE_192000 != 1 << 12 +#if SNDRV_PCM_RATE_5512 != 1 << 0 || SNDRV_PCM_RATE_24000 != 1 << 14 #error "Change this table" #endif
static unsigned int rates[] = { 5512, 8000, 11025, 16000, 22050, 32000, 44100,
48000, 64000, 88200, 96000, 176400, 192000 };
48000, 64000, 88200, 96000, 176400, 192000,12000, 24000 };const struct snd_pcm_hw_constraint_list snd_pcm_known_rates = { .count = ARRAY_SIZE(rates),
~Vinod
On Thu, Jul 25, 2013 at 08:45:15AM +0200, David Henningsson wrote:
On 07/24/2013 07:26 PM, Vinod Koul wrote:
Hey Takashi,
For compressed audio we also need to support the PCM rates of 12 and 24KHz.
Looking at pcm.h these are not defined so we can simply add them at the end. But am worried about wider impact of adding these rates.
Can you let me know if more is required to be done or below is fine
diff --git a/include/sound/pcm.h b/include/sound/pcm.h index 84b10f9..e418d8d 100644 --- a/include/sound/pcm.h +++ b/include/sound/pcm.h @@ -126,6 +126,8 @@ struct snd_pcm_ops { #define SNDRV_PCM_RATE_96000 (1<<10) /* 96000Hz */ #define SNDRV_PCM_RATE_176400 (1<<11) /* 176400Hz */ #define SNDRV_PCM_RATE_192000 (1<<12) /* 192000Hz */ +#define SNDRV_PCM_RATE_12000 <1<<13> /* 12000Hz */ +#define SNDRV_PCM_RATE_24000 <1<<14> /* 24000Hz */
You probably meant (1<<13), not <1<<13> ?
Yup :)
~Vinod
Vinod Koul wrote:
For compressed audio we also need to support the PCM rates of 12 and 24KHz.
Looking at pcm.h these are not defined
You don't need such symbols to use a rate. These symbols are intended for often-used rates. As long as only one or two drivers use a rate, they can just set KNOT and install a constraint.
But am worried about wider impact of adding these rates.
Can you let me know if more is required to be done or below is fine
Grep for 176400 in sound/core/, include/, and in alsa-lib.
Regards, Clemens
On Thu, Jul 25, 2013 at 08:50:40AM +0200, Clemens Ladisch wrote:
Vinod Koul wrote:
For compressed audio we also need to support the PCM rates of 12 and 24KHz.
Looking at pcm.h these are not defined
You don't need such symbols to use a rate. These symbols are intended for often-used rates. As long as only one or two drivers use a rate, they can just set KNOT and install a constraint.
okay so how exactly is the rate passed to driver and converted and sent to drivers?
But am worried about wider impact of adding these rates.
Can you let me know if more is required to be done or below is fine
Grep for 176400 in sound/core/, include/, and in alsa-lib.
SNDRV_PCM_RATE_176400?
~Vinod
--
At Thu, 25 Jul 2013 21:15:01 +0530, Vinod Koul wrote:
On Thu, Jul 25, 2013 at 08:50:40AM +0200, Clemens Ladisch wrote:
Vinod Koul wrote:
For compressed audio we also need to support the PCM rates of 12 and 24KHz.
Looking at pcm.h these are not defined
You don't need such symbols to use a rate. These symbols are intended for often-used rates. As long as only one or two drivers use a rate, they can just set KNOT and install a constraint.
okay so how exactly is the rate passed to driver and converted and sent to drivers?
Define the own hw_constraint. Usually it's implemented via snd_pcm_hw_constraint_list().
But am worried about wider impact of adding these rates.
Can you let me know if more is required to be done or below is fine
Grep for 176400 in sound/core/, include/, and in alsa-lib.
SNDRV_PCM_RATE_176400?
Heh, this was an already added one. Rather take a look at SNDRV_PCM_RATE_KNOT.
Takashi
On Thu, Jul 25, 2013 at 06:31:33PM +0200, Takashi Iwai wrote:
At Thu, 25 Jul 2013 21:15:01 +0530, Vinod Koul wrote:
On Thu, Jul 25, 2013 at 08:50:40AM +0200, Clemens Ladisch wrote:
Vinod Koul wrote:
For compressed audio we also need to support the PCM rates of 12 and 24KHz.
Looking at pcm.h these are not defined
You don't need such symbols to use a rate. These symbols are intended for often-used rates. As long as only one or two drivers use a rate, they can just set KNOT and install a constraint.
okay so how exactly is the rate passed to driver and converted and sent to drivers?
Define the own hw_constraint. Usually it's implemented via snd_pcm_hw_constraint_list().
Thanks
But am worried about wider impact of adding these rates.
Can you let me know if more is required to be done or below is fine
Grep for 176400 in sound/core/, include/, and in alsa-lib.
SNDRV_PCM_RATE_176400?
Heh, this was an already added one. Rather take a look at SNDRV_PCM_RATE_KNOT.
Okay so the orignal problem is that we use SNDRV_PCM_RATE_XXX to send the sample rate to decoder. Some decoders very commonly use 12 and 24Khz, so I would like them to be added.
I think 12 and 24 are fairly common rates and should be added. Do you agree?
~Vinod --
At Fri, 26 Jul 2013 11:39:08 +0530, Vinod Koul wrote:
On Thu, Jul 25, 2013 at 06:31:33PM +0200, Takashi Iwai wrote:
At Thu, 25 Jul 2013 21:15:01 +0530, Vinod Koul wrote:
On Thu, Jul 25, 2013 at 08:50:40AM +0200, Clemens Ladisch wrote:
Vinod Koul wrote:
For compressed audio we also need to support the PCM rates of 12 and 24KHz.
Looking at pcm.h these are not defined
You don't need such symbols to use a rate. These symbols are intended for often-used rates. As long as only one or two drivers use a rate, they can just set KNOT and install a constraint.
okay so how exactly is the rate passed to driver and converted and sent to drivers?
Define the own hw_constraint. Usually it's implemented via snd_pcm_hw_constraint_list().
Thanks
But am worried about wider impact of adding these rates.
Can you let me know if more is required to be done or below is fine
Grep for 176400 in sound/core/, include/, and in alsa-lib.
SNDRV_PCM_RATE_176400?
Heh, this was an already added one. Rather take a look at SNDRV_PCM_RATE_KNOT.
Okay so the orignal problem is that we use SNDRV_PCM_RATE_XXX to send the sample rate to decoder.
Relying only on SNDRV_PCM_RATE_* bits doesn't scale. What if more other rates are needed at the next time?
Some decoders very commonly use 12 and 24Khz, so I would like them to be added.
I think 12 and 24 are fairly common rates and should be added. Do you agree?
It's no solution. It makes things easier, though. Better to fix the code to be able to handle uncommon rates, at first.
Takashi
On Fri, Jul 26, 2013 at 09:09:09AM +0200, Takashi Iwai wrote:
Relying only on SNDRV_PCM_RATE_* bits doesn't scale. What if more other rates are needed at the next time?
Agreed, this is my thinking as well. Remove the usage of macro and use raw rate values...
~Vinod --
On Thu, Jul 25, 2013 at 09:15:01PM +0530, Vinod Koul wrote:
On Thu, Jul 25, 2013 at 08:50:40AM +0200, Clemens Ladisch wrote:
For compressed audio we also need to support the PCM rates of 12 and 24KHz.
Looking at pcm.h these are not defined
You don't need such symbols to use a rate. These symbols are intended for often-used rates. As long as only one or two drivers use a rate, they can just set KNOT and install a constraint.
okay so how exactly is the rate passed to driver and converted and sent to drivers?
Having said that 24kHz sounds awfully familiar...
At Fri, 26 Jul 2013 11:41:03 +0530, Vinod Koul wrote:
On Thu, Jul 25, 2013 at 08:36:44PM +0100, Mark Brown wrote:
Having said that 24kHz sounds awfully familiar...
Yup it is :)
I think todays rates defined in ALSA are HD-Audio based. IIRC HDA doesnt use 12/24 they were not added here!
No, the values were defined at the time of AC97.
Takashi
participants (5)
-
Clemens Ladisch -
David Henningsson -
Mark Brown -
Takashi Iwai -
Vinod Koul