Sampling errors and the influence of processing after sampling
Eugen-Georg Woschni · 2002
As is well-known, it is not possible to fulfil the sampling-theorem exactly: cut-off-errors because of the necessary low-pass-filtering before sampling-to get a band-limited input signal-and if this is not possible, aliasing errors appear. It is shown, that both the cut-off- and aliasing-errors are dependent on the processing algorithm. Processing algorithms with an integral (I-) part, such as typical of mean operations, lead to decreasing errors in comparison with the normally regarded case of no operation after sampling. Algorithms with a differential (D-) part, on the other hand, result in increasing errors. An example showing typical cases using a general signal model of the input is included.