[sdiy] Analysis of frequency variation in analogue synths

[Magnus Danielson]

From: "Paul Perry" <pfperry at melbpc.org.au>
Subject: RE: [sdiy] Analysis of frequency variation in analogue synths
Date: Thu, 3 May 2007 17:49:22 +1000
Message-ID: <005801c78d57$907730b0$8501a8c0 at ice9>

> Oh, I agree entirely.
> I was just saying, that beating with a rock-solid source would at least give
> an indication of longer term variations.
> Anything psychophysically interesting, would presumably involve jitter of
> some kind in the 10mS -500uS range. And time dependent structure in the
> jitter error would be where the effect - if any - would originate.

Actually, I think that you would have to include a larger time-base range since
the ear is able to identify and apprechiate slow shifts in pitch. It is hard to
apriori eliminate a range in time. It is howover customary to measure up to
half the carrier frequency. I would measure to include up to 100 s. The actual
measurement for Allan deviation with tau of 100 s needs to be a multiple longer
naturally, or else the statistical basis for the measure is too low.

Oh, I forgot to mention that Allan deviation and Modified Allan deviation is
needed to handle phase noise which raises with lower frequency, i.e. non-white
noise. Normal standard deviation does not converge for such sources of noises.

Cheers,
Magnus