To desing low pass FIR filter analysis and application

1.to design lp fir filter using fir1, fir2 ,firpm 2. use three cutoff frequencies 500-1000hz, 1001-2000hz,2001-3000hz 3.plot frequency response using freqz and fft 4.find the group delay using angle() unwrap() freqz() and diff() 5.filter signal using filter() and then play original and sound and filtered using soundsc() 6.compare the outputs of myfilter() function to implement filtering operation

amongst all these please let me know how can i use combination like plot ( 20*log10(abs(fft(b,8000)))); type of command..

i already used fir1 and got attenuation of min 40db .. played the original sound signal.. struggling with from 3rd step starting with fft and plot function

thankyou

Answer 1

[Youssef Khmou]

Generally the returned values from filter design function are two vector a and b that represent the nominator and denominator of the the transfer function H, you can follow this protocol :

 % Example of the vector a and b :
  b=[1.2 0.9  1 2 2.3 6];
  a=[1.6 0.75 1 1 -1 0.99]; 
  N=512; % number of points for frequency computation
  Fs=8000;  %  8Khz
  [h,f]=freqz(b,a,N,Fs);
  figure, plot(f,20*log10(abs(h)));

[tilak tenneti]

hi youssef, in this case i have to take filter coefficients for b

i have tried this... x=wavread('thermo.wav'); soundsc(x,8000); b=fir1(100,0.385,'low'); a=fir1(200,0.475,'low'); N=512; Fs=8000; [h,f]=freqz(b,a,N,Fs); figure, plot(f,20*log10(abs(h)));

pls let me know how to use plot with diff angel and unwrap ??

[Youssef Khmou]

You can apply the Discrete Fourier Transform to vectors a and b you created :

 a=fir1(200,0.475,'low');
 Fa=fft(a);
 figure, plot(20*log10(abs(Fa(1:end/2))));
 figure,psd(a), %  Built in function PSD
 % Same goes for b  .....

You can compute the phase angle as the following :

 %Given your signal y original or filtered
 Fy=fft(y);
 P=unwrap(angle(Fy));
 Frequency=(0:length(Fy)-1)'/length(Fy)*Fs;
 figure, plot(Frequency,P);