how to filter a signal

I have a signal and I filtered the signal using a cheby1 filter. the frequencies of the signal range from 0.058 to 349Hz. I had to remove frequencies above 0.7Hz. after filtering the signal again when I find the frequencies I'm getting frequencies above 0.7Hz.

Answer 1

[Star Strider]

I showed you how to correctly design a filter here. If you want to design a filter to remove all frequencies above 0.7 Hz, design a lowpass filter, specify the passband frequency as 0.7/Fn and the stopband at 0.72/Fn. Use a Chebyshev Type II filter for this, instead of a Type I, since you now want a relatively flat passband.

[rohith bharadwaj]

But when I'm using type 2 my filtered signal amplitude is very less. original signal amplitude - 25. filtered - 1*10^-3.

can't I use the cheby1 filter? Is my code wrong?

[rohith bharadwaj]

if my filter code is correct why I am getting frequencies again after filtering. [b,a] = cheby2(1,0.5,0.7/325,'low'); fil=filter(b,a,y);

I've with this. but again when I am finding the frequencies I am getting frequencies above 0.7Hz. Is there any reason for this?

[Star Strider]

Use the design approach I gave you. The second-order-section representation will produce a stable filter.

You haven’t stated your sampling frequency, and there is no time vector in your data file. From your code, I assume it is 650 Hz to produce a Nyauist frequency of 325 Hz.

[rohith bharadwaj]

yes, I've used your approach. My sampling frequency is 650Hz. My output is correct but my filtered signal gain is coming less than that of the original signal. How can I tune the gain or is there any approach to do so.

[Star Strider]

If you used this filter:

[b,a] = cheby2(1,0.5,0.7/325,'low');
fil=filter(b,a,y);

Your filter has one pole, and only has 0.5 dB stopband attenuation.

See if this one works better in your application:

Fs = 650
Fn = Fs/2;                                              % Nyquist Frequency (Hz)
Wp = 0.7/Fn;                                            % Passband Frequency (Normalised)
Ws = 0.72/Fn;                                           % Stopband Frequency (Normalised)
Rp =   1;                                               % Passband Ripple (dB)
Rs =  50;                                               % Stopband Ripple (dB)
[n,Ws] = cheb1ord(Wp,Ws,Rp,Rs);                         % Filter Order
[z,p,k] = cheby1(n,Rs,Ws,'low');                        % Filter Design
[soslp,glp] = zp2sos(z,p,k);                            % Convert To Second-Order-Section For Stability

figure(5)
freqz(soslp, 2^16, Fs)                                  % Filter Bode Plot

[rohith bharadwaj]

thank you :-)

[Star Strider]

My pleasure!

[Wietse van Geel]

EDIT: found your previous post in the link! I'll try that first

Thank you for this helpful post. I'm designing a filter for the first time, and find it quite challenging.

Could you clarify the above answer by explaining how the last part, ending with:

[soslp,glp] = zp2sos(z,p,k);

can be implemented as a filter?

This part works fine for me:

[b,a] = cheby2(1,0.5,0.7/325,'low');

fil=filter(b,a,y);

And playing around with the numbers and plotting the results give some idea of what happens.

Best,

Wietse

[Star Strider]

Wietse van Geel — My pleasure!

I have switched over to elliptic filters, since they are much more computationally efficient and generally perform much better. The code would now be:

Fs = 650
Fn = Fs/2;                                              % Nyquist Frequency (Hz)
Wp = 0.7/Fn;                                            % Passband Frequency (Normalised)
Ws = 0.72/Fn;                                           % Stopband Frequency (Normalised)
Rp =   1;                                               % Passband Ripple (dB)
Rs =  50;                                               % Stopband Ripple (dB)
[n,Wp] = ellipord(Wp,Ws,Rp,Rs);                         % Filter Order
[z,p,k] = ellip(n,Rp,Rs,Wp,'low');                      % Filter Design
[soslp,glp] = zp2sos(z,p,k);                            % Convert To Second-Order-Section For Stability

figure(5)
freqz(soslp, 2^16, Fs)                                  % Filter Bode Plot

Use the filtfilt function to do the actual filtering:

fil = filtfilt(soslp,glp,y);                            % Filter Signal

.

Answer 2

[kani mozhi]

hi... i m wrk in bci data competition iii dataset 1.which filter have to use and please give the matlab code.. i m new in this area