plughw behaviour not consistent with hw

plughw behaviour not consistent with hw

[Stefano Antonelli]

Hello list,

I'm working on an application that requires a period time of 10ms.  The
hardware at the other end of the bus is a ulaw codec running at 8kHz.
 The application runs at 16kHz S16_LE and I was hoping to use the
plughw device to do sample rate conversion and ulaw encode/decode.
 This way the application doesn't need to change to accommodate this
ulaw codec.  However, it's not working and I don't know how to go about
debugging.  I'm hoping I can get some pointers here.

The problem is best illustrated by this output, first using plughw:
alsa_read: 10
alsa_read: 20
alsa_read: 0
alsa_read: 20
alsa_read: 0
alsa_read: 20
alsa_read: 0
alsa_read: 20
alsa_read: 0
alsa_read: 20

It would appear that plughw gives the correct period time on the first
call, but then subsequently buffers 2 periods before returning.  
Using hw with matching sample rate, sample format, and number of
samples:
alsa_read: 10
alsa_read: 10
alsa_read: 10
alsa_read: 10
alsa_read: 10
alsa_r
ead: 10
alsa_read: 10
alsa_read: 10
alsa_read: 10
alsa_read: 10

This code snippet generated the above output:

for (i = 0; i < 10; ++i) {
        gettimeofday(&time1, NULL);
	err = snd_pcm_readi (capture_handle, buf, NSMP); 
        if (err != NSMP) {
                printf ("read from audio interface failed "
                        " (%s)\n", snd_strerror (err));
                exit (1);
        }
        gettimeofday(&time2, NULL);
        printf("alsa_read: %d\n", getTimeDiff(time1, time2));
}

I have tried using the patch from here (without the environment
variable switch):

http://mailman.alsa-project.org/pipermail/alsa-devel/2008-October/01170
3.html

to print out the configuration of plughw's plugin chain, but it
generates no output.  I've verified the md5sum of the library after
compilation and it matches the one I compiled.  And I ran ldconfig
after replacing libasound.so.  So I don't know what's going on there.
 I definitely call snd_pcm_hw_params() when setting up the capture
device.  This is using alsa-lib-1.0.25.

Does anyone know why plughw might be acting the way it is and how I can
get it to deliver 10ms periods?  Equally appreciated would be any tips
on how to debug this problem.

Thanks,
Stef

_______________________________________________
Alsa-devel mailing list
Alsa-devel@alsa-project.org
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

Re: plughw behaviour not consistent with hw

[Takashi Sakamoto]

Hi,

On 2018/11/01 13:44, Stefano Antonelli wrote:
> Does anyone know why plughw might be acting the way it is and how I can
> get it to deliver 10ms periods?  Equally appreciated would be any tips
> on how to debug this problem.

The 'plughw' PCM node of alsa-lib configuration space is assigned to
'plug' PCM plugin[1]. This PCM plugin can aggregate several PCM plugins
to perform resampling. For example, in a below case, two PCM plugins
(rate/route conversion PCM plugins) are aggregated by 'plug' PCM plugin
in front of 'hw' PCM plugin which direct accesses to ALSA PCM chracter
device.

$ LC_ALL=C arecord -v -Dplughw:1,0 /dev/null -fS16_LE
Recording WAVE '/dev/null' : Signed 16 bit Little Endian, Rate 8000 Hz, Mono
Plug PCM: Rate conversion PCM (48000, sformat=S16_LE)
Converter: libspeex (external)
Protocol version: 10002
Its setup is:
   stream       : CAPTURE
   access       : RW_INTERLEAVED
   format       : S16_LE
   ...
Slave: Route conversion PCM (sformat=S16_LE)
   Transformation table:
     0 <- 0*0.5 + 1*0.5
Its setup is:
   stream       : CAPTURE
   access       : MMAP_INTERLEAVED
   format       : S16_LE
   ...
Slave: Hardware PCM card 1 'HDA Intel PCH' device 0 subdevice 0
Its setup is:
   stream       : CAPTURE
   access       : MMAP_INTERLEAVED
   format       : S16_LE
   ...


Well, for capture direction, the most of PCM plugins in
alsa-lib/alsa-utils are programmed to do resampling in .avail_update()
callback of plugin framework in alsa-lib. The number of available PCM
frames is calculated in this timing. On the other hand, poll(2) system
call is executed against ALSA PCM character device. Even if poll(2)
returns to notify available PCM frames in the device, actual number of
available frames is a summary of resampling done by the
aggregated PCM plugins. As a result, you can see the behaviour because
plugins possibly have buffers for count gap between in/out frames.

[1] 
http://git.alsa-project.org/?p=alsa-lib.git;a=blob;f=src/conf/alsa.conf;h=bb00ff409cd9643b01e2d3f8e36dc9571ddb0c85;hb=HEAD#l193


Regards

Takashi Sakamoto
_______________________________________________
Alsa-devel mailing list
Alsa-devel@alsa-project.org
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel