A unified approach to the restoration of lost samples in discrete-time signals.

The problem of estimating lost sample values in discrete-time signals is considered. The problem is treated as a linear minimum variance estimation problem, which, in principle, requires knowledge of the signal's autocorrelation coefficients. It is shown that, starting from this general principle, several sample restoration methods studied separately in the literature can be derived. As particular examples, the authors consider sample restoration methods for bandlimited signals, multiple sinusoids, autoregressive processes and (quasiperiodic) speech signals.