An approximately unbiased recovery of discrete Fourier transforms altered by jittered sampling epochs

The effects of a time jitter on the computed D.F.T. of analog signals are investigated. It is shown that it results in a bias described as a filtering effect based on the knowledge of the characteristic function of the jitter. An inverse filtering is proposed. With the additional hypothesis of jitter sample to sample independent, the computations are also carried out for squared modulus of D.F.T., or P.S.D. of stochastic processes. Simulation results are given, showing good agreement with theoretical expressions. Application fields include debiased recovery of true D.F.T., as well as jitter characterization.