The applet designs an FIR filter by the Window method, based on the Kaiser window function. The applet determines the required values of the filter order and Kaiser window parameter (which controls the shape of the window function) to achieve the transition bandwidth and stopband attenuation specified by the user. Once the filter design has been carried out, the frequency response of the filter can be plotted and the filter coefficients listed.
Note that the Kaiser window function differs from fixed-shape window functions such as the rectangular window and the Hanning window in that its shape is continuously variable to provide essentially any required amount of stopband attenuation. However, improved stopband performance is achieved at the expense of increased transition bandwidth, and this can be reduced only by increasing the filter order.
The filter design is based on a sampling rate of 8000 samples/s.
Select a low pass, high pass or band pass filter and enter values for:
Click the "Design filter" button. The order of filter satisfying these requirements is shown.
Now click the "Frequency response" button. The gain (magnitude) frequency response of the filter is calculated at uniform intervals across the frequency range from 0 to 4000 Hz (half the sampling rate). Note that it may take some time for the filter design and frequency response calculations to be completed on slower machines - please be patient!
The frequency (horizontal) scale is divided into intervals of 500 Hz. The gain (vertical) scale extends from 0 dB (unit gain) at the top to -100 dB in steps of 10 dB.
Click the Coefficients button to list the filter coefficients. The coefficients are normalised to make the maximum filter gain equal to 1. If you wish, you can copy and paste the coefficient list to a text editor; this will enable you to save the coefficients to disc. (A Java applet isn't allowed to write to your local drive).
Download source code (zip Archive)
FIR filter design applet (based on fixed window functions): Java 1.02 version | Java 1.1 version