[alsa-devel] [ISSUE] rtl8192ce appears to interfere with ALSA playback

[Takashi Iwai]

At Thu, 23 Jan 2014 16:36:45 +0000,
LANGLOIS Olivier PIS -EXT wrote:
> 
> > > As small unrelated sidenote,
> > >
> > > If I let aplay choose the hw params itself, how comes it ends up with 16
> > periods when it calculates the period time to have 4 periods?
> >
> > How 4 periods come up...?
> > period_time = period_size / rate.
> >
> >
> I was referring to this code in aplay.c:set_params():
> 
>         if (period_time == 0 && period_frames == 0) {
>                 if (buffer_time > 0)
>                         period_time = buffer_time / 4;
>                 else
>                         period_frames = buffer_frames / 4;
>         }

But you passed --buffer-size and --period-size options.
> 
> and I think that usually periods_num ~= buffer_time/period_time or buffer_size/period_size
> 
> aplay -Dhw:0,0 sine_48000.wav
> 
> I have:
> 
>   buffer_size  : 24000
>   period_size  : 6000
>   period_time  : 125000
> 
> but with the default dmix setup: aplay -v sine_48000.wav:
> 
>   buffer_size  : 16384
>   period_size  : 1024
>   period_time  : 21333

As already mentioned, dmix has the fixed setup for its slave PCM.
See src/conf/pcm/dmix.conf.


Takashi