[ntp:questions] Accuracy of audio tones via VOIP

unruh unruh at invalid.ca
Wed Jul 17 16:33:27 UTC 2013

On 2013-07-17, Robert Scott <no-one at notreal.invalid> wrote:
> On Tue, 16 Jul 2013 15:34:58 GMT, unruh <unruh at invalid.ca> wrote:
>>Now, I strongly suspect that his competition, who claim 10PPM accuracy
>>on their ability to measure the frequency (against an absolute time
>>standard) of a piano string, are just lying.
> I never said they could measure the frequency of a piano string to
> that accuracy.  I said that their basic calibration was accurate to
> 10ppm.  So if you used their devices to measure a stable tone from a
> laboratory audio signal generator, you could measure it with that much
> accuracy.  You are right that real-world piano strings are not so
> stable and have too short a sustain to measure any better than about
> 0.2 cents (about 115ppm).  But in keeping with long-established

Sorry, I do not believe .2 cents either. even 1 cent I would regard as
maybe possible, but probably a stretch.
I looked at the web page of verituner(?), and they  list their frequencies
to .01 cents-- even the higher harmonics (which die out even faster) which is rediculous.

I suspect that the pitch of the string varies by more than that as a
function of the amplitude of oscillation of the string. And certainly,
the pitch change, when you go from one string to two or three strings
playing in unison, is more than that (soundboard loading of the string).
And the psychoacoustics of hearing the pitch of a note whose
partials are not harmonics is a can of worms.

> principles in scientific measurement, it is good to have instruments
> that are at least an order of magnitude more accurate than inherent
> inaccuracies of the thing you are trying to measure.  That way to
> don't start unnecessarily giving up margin.
> One on my competitors actually delivers a physical tone generator
> along with the software purchase.  It is just a cheap commercial
> musical pitch reference that they measure and document against a
> laboratory standard before they send it out.

Except, as you know (or should know if you have followed this newsgroup
for any length of time) the crystal oscillator in that tone generator
will drift, and after a few months may certainly be out by more than
10PPM. Since it is also not a temperature controlled oscillator, it will
also depend on the temperature that the crystal is at, etc. So he may
claim that he calibrated it to 10PPM, but not mention that what the
person receives and especially what he uses is not that accurate. 
>>But that is the challenge he has set himself. Note that he wants
>>his clients to connect to that tone and to have that tone delivered
>>reliably, for many minutes, in order to calibrate to that accuracy
> I can achieve the desired 10ppm lock on the NIST tones in about 30
> seconds if the audio is clear.  It is not hard.

In 30 sec a 10PPm lock is a phase accuracy of 6 degrees which while
certainly not impossible, is, in the presense of the noise,etc on the
phone line pusing it. 

> Robert Scott
> Hopkins, MN

More information about the questions mailing list