Butterworth filter

Prev: inverse prediction for logistic regression?
Next: colorbar title

From: daffidity Ming on 8 Jul 2010 17:31

---

Hi,

I am new to Matlab and am trying to write a butterworth filter. Matlab help and Google gave me some indication as to what the code should look like but I still dont understand why it does not work. I am attempting to use a band pass filter to remove 1 frequency from a mixture of sine waves.

% bandpass filter on a set of sin waves
clear all;
figure(1); clf;

x = 0:0.1:500;

y = 1.5*sin(2*pi*x/100) + 1*sin(2*pi*x/20); % frequencies present = 0.01 0.05 Hz
plot(x,y, 'g');

f_nyquist = 2*pi/(0.1);
n = 5
cutoff_f = [0.03 0.07]/f_nyquist %I would like the extract the 0.05 Hz sine wave

[b, a] = butter(n, cutoff_f, 'bandpass');

output = filter(b, a, y);
figure(3); clf;
plot(x, output);

Any help would be appreciated.

D

From: Wayne King on 8 Jul 2010 18:07

---

"daffidity Ming" <adona13.alice(a)gmail.com> wrote in message <i15g2o$gg5$1(a)fred.mathworks.com>...
> Hi,
>
> I am new to Matlab and am trying to write a butterworth filter. Matlab help and Google gave me some indication as to what the code should look like but I still dont understand why it does not work. I am attempting to use a band pass filter to remove 1 frequency from a mixture of sine waves.
>
> % bandpass filter on a set of sin waves
> clear all;
> figure(1); clf;
>
> x = 0:0.1:500;
>
> y = 1.5*sin(2*pi*x/100) + 1*sin(2*pi*x/20); % frequencies present = 0.01 0.05 Hz
> plot(x,y, 'g');
>
> f_nyquist = 2*pi/(0.1);
> n = 5
> cutoff_f = [0.03 0.07]/f_nyquist %I would like the extract the 0.05 Hz sine wave
>
> [b, a] = butter(n, cutoff_f, 'bandpass');
>
> output = filter(b, a, y);
> figure(3); clf;
> plot(x, output);
>
> Any help would be appreciated.
>
> D

Hi D,

How about using filter objects.

x = 0:0.1:500;
y = 1.5*sin(2*pi*x/100) + 1*sin(2*pi*x/20);
% I'm just using your parameters--not saying it's the best filter
h = fdesign.bandpass('N,F3dB1,F3dB2',10,0.03,0.07,10);
D = design(h,'butter');
output = filter(D, y);
plot(x,output);

Hope that helps,
Wayne

From: daffidity Ming on 8 Jul 2010 19:36

---

Thank you for pointing me in the right direction! it works and I am reading the documentation on it.


"Wayne King" <wmkingty(a)gmail.com> wrote in message <i15i68$1t2$1(a)fred.mathworks.com>...
> "daffidity Ming" <adona13.alice(a)gmail.com> wrote in message <i15g2o$gg5$1(a)fred.mathworks.com>...
> > Hi,
> >
> > I am new to Matlab and am trying to write a butterworth filter. Matlab help and Google gave me some indication as to what the code should look like but I still dont understand why it does not work. I am attempting to use a band pass filter to remove 1 frequency from a mixture of sine waves.
> >
> > % bandpass filter on a set of sin waves
> > clear all;
> > figure(1); clf;
> >
> > x = 0:0.1:500;
> >
> > y = 1.5*sin(2*pi*x/100) + 1*sin(2*pi*x/20); % frequencies present = 0.01 0.05 Hz
> > plot(x,y, 'g');
> >
> > f_nyquist = 2*pi/(0.1);
> > n = 5
> > cutoff_f = [0.03 0.07]/f_nyquist %I would like the extract the 0.05 Hz sine wave
> >
> > [b, a] = butter(n, cutoff_f, 'bandpass');
> >
> > output = filter(b, a, y);
> > figure(3); clf;
> > plot(x, output);
> >
> > Any help would be appreciated.
> >
> > D
>
> Hi D,
>
> How about using filter objects.
>
> x = 0:0.1:500;
> y = 1.5*sin(2*pi*x/100) + 1*sin(2*pi*x/20);
> % I'm just using your parameters--not saying it's the best filter
> h = fdesign.bandpass('N,F3dB1,F3dB2',10,0.03,0.07,10);
> D = design(h,'butter');
> output = filter(D, y);
> plot(x,output);
>
> Hope that helps,
> Wayne

 | 
Pages: 1
Prev: inverse prediction for logistic regression?
Next: colorbar title