[time-nuts] Oh dear

[Chris Albertson]

New question about "jitter" in recording.    I was reading some time
ago about non-uniform sampling.  Basically the time between samples is
random (or as random as you can make it)  But now you have to sample a
clock AND the signal. Or more likely use a psuedorandon sample
interval that can be reconstructed without clock samples     The main
problem with this technique is that few people understand the math.
For example what is the frequency response of a system with a given
mean and standard deviation sample period?  Advantages are that you
can sample higher frequency than 1/2 the average sample rate and
alieasing is less a problem.


Chris Albertson
Redondo Beach, California