<p><br>
><br>
> > How accurate are the hwptr playback/capture position ?<br>
>><br>
>><br>
>> e,g, USB audio send in URB packet<br>
>><br>
>> the position is only updated in one ms<br>
><br>
><br>
> There is a chance to update the position every 2 msec. But it differ depending on sampling rate, size of period and streaming mode.<br>
></p>
<p>Do you mean the card have internal memory of FIFO to hold 1 packet of data ?</p>
<p>Can the driver run with one periods per buffer or need two or more periods ?</p>
<p>Why periods_min = 1 ?</p>
<p><a href="http://www.freedesktop.org/wiki/Software/PulseAudio/Backends/ALSA/BrokenDrivers/">http://www.freedesktop.org/wiki/Software/PulseAudio/Backends/ALSA/BrokenDrivers/</a></p>
<p>><br>
><br>
>> Do you mean firewire driver 's pcm_pointer callback does not report<br>
>> playback/capture positon in realtime ?<br>
><br>
><br>
> Within the 2 msec, I mean it.<br>
><br>
><br>
>> there is alway one packet 's uncertainty<br>
><br>
> ><br>
> > are the any minimum size of the packet ?<br>
> ><br>
> > e.g. usb audio send in urb<br>
><br>
> There is uncertainty. But not one packet.<br>
><br>
> Every 2 msec, the drivers queue some packets for transferring. So uncertainty is the number of the packets.<br>
><br>
> Currently this number is fixed by 16. See 'INTERRUPT_INTERVAL' macro in 'amdtp.c'.<br>
> (In IEE1394, 8000 packets are transferred per sec. So 16 packets equal 2 msec.)<br>
><br>
></p>
<p>Do the driver need to use larger value for pulseaudio's rewind_safeguard as it use more channels than the other cards ?</p>
<p><a href="https://bugs.freedesktop.org/show_bug.cgi?id=72396">https://bugs.freedesktop.org/show_bug.cgi?id=72396</a><br></p>