Scribd
Upload
300 views

DSP Cep

The document describes a group project on designing digital filters to remove noise from a signal. The group members are Arfan Shahzad, Nouman Saeed, and Tehzib-ul-Hasnain. They discuss designing finite impulse response (FIR) and infinite impulse response (IIR) filters using methods like frequency sampling, windowing, and bilinear transformation. Parameters like cutoff frequencies, passband ripples, and transition bandwidth are specified. FIR filters are designed using windowing methods and IIR filters using Butterworth filters. The filters are applied to a noisy signal to generate a filtered output.

Uploaded by

Kashif Hassan
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
300 views

DSP Cep

The document describes a group project on designing digital filters to remove noise from a signal. The group members are Arfan Shahzad, Nouman Saeed, and Tehzib-ul-Hasnain. They discuss designing finite impulse response (FIR) and infinite impulse response (IIR) filters using methods like frequency sampling, windowing, and bilinear transformation. Parameters like cutoff frequencies, passband ripples, and transition bandwidth are specified. FIR filters are designed using windowing methods and IIR filters using Butterworth filters. The filters are applied to a noisy signal to generate a filtered output.

Uploaded by

Kashif Hassan
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 15

DIGITAL SIGNALS PROCESSING

Complex Engineering Problem

Design of FIR and IIR Filters


Group Members

ARFAN SHAHZAD 15-EE-100


NOUMAN SAEED 15-EE-104
TEHZIB-ul-HASNAIN 15-EE-184
Filter Design

FIR (finite impulse response)


 Frequency sampling method
 Windowing method
Filter Design

IIR (infinite impulse response)


 By impulse invariance method
 Bilinear transformation method
Characteristics

 FIR can achieve desirable linear phase response


 FIR are inherently stable (absolutely summable)
 IIR has low order (lower complexity)
 IIR needs attenuation for stability
FILTER DESIGN PARAMETERS

 Stop band frequency (Ws) ([415,425],[745,755] FIR)


([410,430],[740,760] IIR)
 Pass band frequency (Wp) ([410,430],[740,760] FIR)
([370,470],[700,800] IIR)
 Stop band attenuation (As) (-60 dB)
 Pass band Ripples (Ap) (0.1dB)
 Transition Bandwidth (60 Hz FIR) (40 Hz IIR)
FIR DESIGN BY WINDOWING METHOD

 From given specification parameters determine the ripples $s and $p


as
Ap = 20log10[(1+$p)/(1-$p)]= 0.1dB; $p = 0.0058
As = -20log10($s) = -60dB $s = 0.0010
 Determine the cut off frequency as
Wc = (Ws+Wp)/2 = ([370 470], [700,800] FIR)
([390,450], [720,780] IIR
 Determine the design parameter as
A=-20log10($)=As=60dB
Transition bandwidth= deltaW=Ws- Wp=([60Hz FIR], [40Hz IIR])
FIR DESIGN BY WINDOWING METHOD

 Select the approximate window of desired A


and find the order of filter
M=L-1
Blackman Window having As of 60dB
Determine the impulse response using filter
commands i.e. FIR1 and FIR2
 Check whether the design filter satisfies the
prescribed specification, if not increase the
order of filter
PLOT NOISY SIGNAL

load('noisy_signal.mat')
y = (noisy_signal);
time = 10;
[l w] = size(noisy_signal);
Fs = l/time;
n = [0:1:length(noisy_signal)-
1]*1/Fs;
figure (1)
plot(n,y)

title('Noisy Signal');
xlabel('Time (s)');
ylabel('Amplitude');
//Fourier Transform of Noisy Sig

NFFT_input =
2^nextpow2(L_input)
Y = fft(input,NFFT_input)/L_input;
f=
Fs/2*linspace(0,1,NFFT_input/2+1)
;
plot(f,(abs(Y(1:NFFT_input/2+1))))
xlabel('Frequency(Hz)')
ylabel('Magnitude')
title('Spectrum of Noisy Signal')
Now to find the frequencies of noise

noise = input(60000:80000); // here it is pure noise


//F.T. OF Noise
L_noise = length (noise);
NFFT_noise = 2^nextpow2(L_noise)
Y = fft(noise,NFFT_noise)/L_noise;
f = Fs/2*linspace(0,1,NFFT_noise/2+1);
plot(f,20*log10(abs(Y(1:NFFT_noise/2+1))))
xlabel('Frequency(Hz)')
ylabel('Magnitude')
title('Spectrum of Noise’ )
Noise Frequencies : 420Hz and 750Hz with max amplitude of 0.025
FIR design using fir1() function
wp_L = [370/4000 , 470/4000] ;
wp_H = [700/4000 , 800/4000] ;

fir_L = fir1(412,wp_L,'stop',hann(413));
fir_H = fir1(412,wp_H,'stop',hann(413));
fir_filter = conv(fir_L,fir_H);

filtered_signal=conv( fir_filter,input);
IIR design using Butter() and Butterord()
Ap = 0.1; As = 32;
wp1 = 370/4000; wp2 = 470/4000 ; ws1 = 410/4000; ws2 = 430/4000
;
wp3 = 700/4000; wp4 = 800/4000 ; ws3 = 740/4000; ws4 = 760/4000
;
[NL,WcL] = buttord([wp1 wp2],[ws1 ws2],Ap,As)
[BL,AL] = butter(NL,WcL,'stop’);

[NH,WcH] = buttord([wp3 wp4],[ws3 ws4],Ap,As);


[BH,AH] = butter(NH,WcH,'stop’);

Bt=conv(BL,BH);
At=conv(AL,AH);
filtered_signal=filter( Bt,At,noisy_signal);

You might also like