[alsa-devel] snd-usb-audio syncronisation problem
Adrian Stancu
adrian.stancu.82 at gmail.com
Fri Feb 22 14:45:07 CET 2008
Hi Clemens,
Thank you very much for helping me!
Using the original module, the momentary frequency is always stalled
at 48000Hz during playback:
vader:~# while sleep 0.3; do grep "Momentary freq"
/proc/asound/card1/stream0; done
Momentary freq = 48000 Hz (0x6.0000)
Momentary freq = 48000 Hz (0x6.0000)
Momentary freq = 48000 Hz (0x6.0000)
Momentary freq = 48000 Hz (0x6.0000)
Momentary freq = 48000 Hz (0x6.0000)
note: I started playback after issuing the command.
So maybe this particular type of interface does not report back
correctly, thus confusing the driver?
After playing with subs->freqm in init_substream_urbs(,,,), the
situation improved dramatically for me, but this is no general fix...
In this case, I get:
Momentary freq = 48001 Hz (0x6.0007) and I get a single little pop
once about 90 seconds, but this is nothing compared to the horrible
sound I normally get when the buffer gets empty.
On Fri, Feb 22, 2008 at 3:09 PM, Clemens Ladisch <clemens at ladisch.de> wrote:
> Adrian Stancu wrote:
> > I'm using an E-MU 0404 USB interface with the snd-usb-audio module
> > from alsa 1.0.16, operating in isochronous asynchronous mode. This
> > interface was producing very annoying click/pop sounds at regular
> > intervals during playback. Eventually I discovered that the
> > syncronisation mechanism telling the driver to send audio data faster
> > or slower to match the interface's consumption speed does not work as
> > expected, for some reason.
>
> The momentary frequency as requested by the device is shown in
> /proc/asound/card?/stream? (if CONFIG_SND_VERBOSE_PROCFS is enabled).
>
> Try something like:
>
> while sleep 0.3; do grep "Momentary freq" /proc/asound/card0/stream0; done
>
> In theory, once the device has determined its own speed relative to
> the computer, this value should be nearly constant. (When a stream has
> just been started, the requested frequency usually is higher because the
> device's buffer is empty.)
>
>
> Regards,
> Clemens
>
More information about the Alsa-devel
mailing list