Dealing with Random Errors in Sampling Rates?

Question

I wonder what generalization I can make about what happens when I introduce some error in my sampler spacing.

For example, If I sample once every second I would want my sampler to take a signal at 1s,2s,3s,4s.

Unfortunately, my real world sampler looks like more like 1.01s, 1.99s, 2.99s, 4.01s.

I want to make a statement about the Fourier domain. I expect my high frequencies to be ambiguous while my lower ones should not to change, too much. Is this correct?

I understand how to do this in terms of a uniform change, for example 1.01, 2.02, 3.03 (which is just a change in sampling rate), but I'm wondering how I can describe this with some knowledge of my error distribution, or for my application even understanding how to describe this knowing the mean of the error would be wonderful.

Can somebody point me in the right direction?

Answer 1

You might read up on "Aperture Jitter". You are correct that higher frequencies are affected more than lower ones. Here is a good reference:

http://www.analog.com/static/imported-files/tutorials/MT-007.pdf