FIR Filter Design Choices

The parameters that we use to design an FIR filter depend on the system that it is going to be implemented on. Some of the parameters that we would like to choose when designing an FIR filter are the sampling frequency, the filter length, frequency response, filter type, windowing function, and time domain properties. We would like the sampling frequency to be as fast as possible. A longer filter will give more resolution in the frequency domain. Input signals may have important time-domain information that we do not want to distort or corrupt. The filter length depends on the performance of the hardware that the filter is being implemented on. A large filter length will be prohibitively slow. The sampling frequency of the filter must be less than the Nyquist limit. FIR filter taps are coefficient/delay pairs. The number of filter taps indicates the memory needed to implement the filter, the number of calculations that are required to perform in the filter, and the amount of filtering that the filter can accomplish. More taps mean that there is more stop-band attenuation, less ripple, and a narrower frequency response.

[[Hamming Window Function]] – good choice for window function.

Sources

Backlinks

FIR Filters
[[Nyquist Frequency Limit]]
[[Sampling Frequency]]
[[Window Function]]