What is an anti-aliasing filter?
The anti-aliasing filter (AAF) is essentially an insurance policy. If you run a finite sample rate when digitizing data, you have to be on guard against signals of appreciable amplitude that have frequency content of 1/2 the sample rate of greater. Since it takes a minimum of two data points to reconstruct the frequency content of a waveform, signals with frequency content greater than half the sample rate cannot be reconstructed. These “aliased” signals fold back to form erroneous signals with lower frequency attributes. Once aliased data enters the digitzed data stream, there is no way to remove it, or tell it from real data. Thus, filters are used to restrict the bandwidth of the data acquisition front-end. This reduces the possibility of aliased data. Most anti-aliasing filters are chosen at about 1/3 the intended sample rate, as a minimum.
