OVERVIEW

Digital filters are classified as IIR, or Infinite Impulse Response, filters and FIR, or Finite Impulse Response filters. In IIR filters, the previous out- and input are fed back to the system after multiplying with a complex quantity, whereas in the FIR filters, only the previous input is used in the feedback part, thus providing a linear phase response. We have discussed the IIR filters in the previous chapter and designed a few, including the high-pass, low-pass, band-pass, and band-stop filters. In all, the computation time was fast, as it did not take much to calculate a few poles and zeros, but the phase response was generally poor. The initial transient response in IIR filters lingers on for a long time. The filters require values from the past output, and some parts of the initial values are always present in the current output, in theory at least. Though, in practice, the fraction grows less and less and eventually reaches 0, it might take a very long time before the floatingpoint arithmetic precision gives way to the accuracy of the calculated numbers. The linear phase response makes the FIR filters suitable candidates where the phase information has an important role to play.

We now discuss a different method of obtaining a filter Transfer Function that would simply rely upon the values from the past input but not the past outputs, as such; there is a uniform phase delay with no distortion. The linear phase is a definite plus, where phase...