## problem with filtering a signal

### hossein (view profile)

on 8 Jul 2013

I wanna filter a signal by band-pass filter and the problem is when I exert the filter on my signal the output is so wrong.I try to exert the filter on a easy signal to test it: for example when I exert a band pass filter with 8-12 frequency range to the x=sin(2*pi*10*t)+sin(2*pi*20*t) the output must similar to sin(2*pi*10*t).but its not. I dont know where I am wrong?

### Wayne King (view profile)

on 8 Jul 2013

```    Fs = 250;
t = 0:1/Fs:10-1/Fs;
x = sin(2*pi*10*t)+sin(2*pi*20*t)+0.25*randn(size(t));
y = sosfilt(SOS,x);```

I used the following parameters in fdatool:

Fs = 250 fstop1 = 7 fpass1 = 9 fpass2 = 12 fstop2 = 14

60 dB of attenuation in both stopbands and 1 db of passband ripple.

hossein

### hossein (view profile)

on 9 Jul 2013

so what about the gain of filter. the y has the amplitude by 10^4 times bigger than x here. you didn't use the filter's gain

### Wayne King (view profile)

on 8 Jul 2013
Edited by Wayne King

### Wayne King (view profile)

on 8 Jul 2013

You need to post the code you have used to design your filter. Also, include the sampling rate for your data.

For example:

```    Fs = 100;
t = 0:1/Fs:10-1/Fs;
x = sin(2*pi*10*t)+sin(2*pi*20*t)+0.25*randn(size(t));
d = fdesign.bandpass('Fst1,Fp1,Fp2,Fst2,Ast1,Ap,Ast2',4,8,12,16,50,1,50,100);
Hd = design(d);
y = filter(Hd,x);
periodogram(y,[],length(y),100)```

hossein

### hossein (view profile)

on 8 Jul 2013

I design the filter by fdatool by these specification:

chebyshef type 2; fs=250; fstop1=7;fpass1=9;fpass2=12;fstop2=13;

then I get the SOS and G matrixes then

``` [a,b]= sos2tf(SOS,G)
x= sin(2*pi*10*t)+sin(2*pi*20*t)
y=filtfilt(a,b,x) or y=filter(a,b,x)```