Minimal one feedback is needed, and is recursive filters. Transfer is done by quotient of polynoms. Zeros realize nonrecursive part, poles recursive part. Order of filter is done by higher of polynoms order. Order is much lower than FIR filters, so response is faster. Example of simplified IIR filter is in follow picture.
Filters are also divided by purposes:
1. Fervency selective digital filters
This is Lowpass (DP) , highpass (HP) pass filter (PP) and
Band-stop filter (PZ).
Realized as FIR or IIR filters with linear phase and as FIR filters.
2. Discrete integrator
Integral is computed by numerical procedures like e.g. y(n) = y(n-1) + x(n)
3. Discrete differentiator
Differentiate input signal, important for acceleration detection or edge detections.
4. Moving average MA
Computing average from sample and M previous. Can weighted with same coefficients
,(no weighted), or with exponential forgetting as recursive filter with weights of coefficients a-ci, where c is real number.