[ntp:questions] Using NTP to calibrate sound app

no-one at no-place.org no-one at no-place.org
Mon Jan 28 20:11:46 UTC 2013

On 28 Jan 2013 17:56:10 GMT, Rob <nomail at example.com> wrote:

>In my opinion, when you think a crystal is not accurate enough, relying
>on single-shot time measurements via radio internet connection will
>*certainly* not be accurate enough.  The protocols used on radio
>channels are not symmetric, because the topology is not symmetric.
>(there is a single base station communicating with a number of clients)
>Typical crystal accuracy is in the same ballpark as what you require.
>When you see significant differences, it is more likely that the actually
>used crystal frequency is different from what you are reading from
>some system information call.

Yes, that may be true.  Smartphone manufactures, for whatever reason,
sometimes design a sound system that delivers 22,150 samples per
second when my app requests that more standard 22,050.  By the way,
this has not been a problem with Apple devices (iPhone/iPad), which
all seem to be close enough to nominal without custom calibration.
But Android devices and Windows laptops are a different story.

It is also my experience that long term crystal stability in these
devices is way better than their absolute accuracy.  So a one-time
calibration is a big improvement over assuming the nominal sampling
frequency.  As I have pointed out before, whatever the timing jitter
is in a radio internet connection, a sufficiently long calibration run
with a sample at the beginning and at the end can give me the required
.02 cent accuracy  (12 PPM).  And there is every indication that the
required time period will be less than the time it takes the user to
get a good night's sleep.

Robert Scott
Hopkins, MN

More information about the questions mailing list