Bug#723924: Update to the bugreport of periodic choppiness in the received audio with Linphone

[Vladimir Chernyshov]

I am not absolutely sure now, that the problem depends on the soundcard 
oscillators. It may be possibly triggered by RX packet loss too.

It can be observed, that the fullness of the output buffer in ALSA 
gradually decreases during a call, and eventually the buffer underflows, 
which is exactly when audio becomes choppy:

In the beginning of a call:
$ cat /proc/asound/card0/pcm0p/sub0/status

state: RUNNING
owner_pid   : 20269
trigger_time: 34665.978754170
tstamp      : 34668.716313151
delay       : 464
avail       : 1584
avail_max   : 1712
-----
hw_ptr      : 21936
appl_ptr    : 22400

During the first attack of the problem:

state: RUNNING
owner_pid   : 20269
trigger_time: 34665.978754170
tstamp      : 35142.111716824
delay       : 0
avail       : 2048
avail_max   : 2112
-----
hw_ptr      : 3809920
appl_ptr    : 3809920

After the first attack of the problem subsides:

state: RUNNING
owner_pid   : 20269
trigger_time: 34665.978754170
tstamp      : 35313.381090594
delay       : -96
avail       : 2144
avail_max   : 2272
-----
hw_ptr      : 5180384
appl_ptr    : 5180288

As the buffer is circular, the problem reoccurs every time the underflow 
delay breaks through an integer multiple of the buffer's length.

Best regards,
Vladimir