Digital Signal Processing Reference
In-Depth Information
FIGURE E.4. Frequency response of FIR bandpass filter using coefficient file BP43K.cof
generated with filter package on CD.
2. Enter the specifications shown in Figure E.3 c . Choose the float option (Figure
E.3 d ) to save the 43 resulting coefficients into a file in a float format (the fixed
option saves the coefficients in hexadecimal). Save it as BP43K.cof.
3. Edit it (an edited version is on the CD). Include it in the program FIRPRN.c
in Example 4.4. Build/run and verify the frequency response of the FIR
bandpass filter centered at 1000 Hz shown in Figure E.4, obtained with an HP
analyzer. An internally generated noise sequence becomes the input to the
FIR filter in the program FIRPRN.c. This filter was designed so that the
center frequency is at 1000 Hz with a sampling frequency of 8 kHz.
E.3.2 Hamming Window
Repeat this procedure for a Hamming window function. Enter 900 and 1100 for the
lower and upper cutoff frequencies. Enter 5.2 (ms) for the duration D of the impulse
response, since the number of coefficients N is
(
) + 1
ND s
This will yield a design with 53 coefficients. Save the resulting coefficient file as
BP53H.cof. Edit it as with the Kaiser window, test it using the program
FIRPRN.C , and verify an FIR bandpass filter with a narrower mainlobe.
Search WWH ::




Custom Search