A filter is basically a system which allows a
particular band of frequencies to pass and selectively changes the wave shape,
phase- frequency characteristics in a desired fashion.In this method the
given frequency response is sampled as a set of equally sampled frequencies to
obtain N samples.The continuous frequency response Hd(w) at N points
essentially gives us the N-Point DFT of Hd(2pnk/N).Thus by using IDFT formula,
the filter co-efficients can be calculated. Increasing the number of samples
can possibly reduce the error. Thus FIR design was implemented using FSM
method.
Simplest method for filter design.
ReplyDeleteFSM realization is more efficient than FIR filter realization.
ReplyDeleteFSM is used in the cases where desired response in known.
ReplyDeleteFSM is also the most simplest and direct technique.
ReplyDeleteDiscontinuity is observed in phase plot between lobes and also when the phase spectrum goes out of the range of -pi to pi.
ReplyDeleteFSM is very efficient
ReplyDeleteThis method requires less computations than windowing method.
DeleteThis method requires less computations than windowing method.
DeleteAs frequency samples increases,error reduces.
ReplyDeleteTo put that in other words your 'N' that is the number of samples of your frequency response should be higher for an improved resolution and reduced error.
DeleteGood for narrow band design
ReplyDeleteFSM is better
ReplyDeleteCompared to windowing method, Yes !
Delete