How to use low pass flter for time domine data

Hello

I need to filter high frequency data and allow only low frequency.

Can I use low pass filter in time domine data to filter noise signal without using FFT to convert data in frequency domine and use a low pass filter then take ifft.

Pls some one help me.

--Thanks

Answer 1

[Youssef Khmou]

hi, you can use built-in functions "butter" and " filter" , example :

 Fs=8000; % SAMPLING FREQUENCY
 t=0:1/Fs:2;
 f1=500; %lower frequency
 f2=1500; % upper frequency
 y=chirp(t,f1,2,f2); % THE SIGNAL WITH BANDWIDTH 1000 Hz
 Fc=1500; %  cutoff frequency .
 Fn=Fs/2;
 % Filter implementation
 [b,a]=butter(2,Fc/Fn,'low');
 x=filter(b,a,y);

I hope this helps

[Thulasi Durai Durai Samy]

Hi Thanks for your reply

My Frequency specification

Sampling Frequency: 500 Hz

Block size: 4500

Therefore total time: 9 seconds

The data has range of Min: 2 Hz to Max: 500 Hz

I need a cut off frequency of 20 HZ

Pls let me know accordingly how to use the above command it will be soooo great full.

------ Thanks

[Thulasi Durai Durai Samy]

Hello Sir

How to use this filter for the time domine data which I have

[Youssef Khmou]

replace the values :

 Fs=500;
 t=0:1/Fs:9-1/Fs; % 4500 points
 f1=2;
 f2=500;
 y=chirp(t,f1,1,f2,'linear'); % the modulation is linear 
 fc=20; 
 [b,a]=butter(2,fc/Fs/2,'low');
 x=filter(b,a,y);

concerning the first signal y, you cant of course visualize the FFT because FS must be > 2* Max(F) , at least 1000 Hz, but the new signal x can be visualized in Freq domain .

[Thulasi Durai Durai Samy]

Sir,

Thanks for your reply,

I have a few questions? pls reply

My data is in time domine - acceleration(y axis) vs time (x axis),

I know only the sampling frequency as 500 Hz and block size as 4500 data points.

1) How do I decided f1 and f2, of course fc is the cut off frequency

2) What is the signification of 'x' data, do I need to multiply 'x' data with my time domine data and see the plot. or I should take fft of my time domine data and then multiply with 'x' data.

My aim is to remove noise signal from the time domine data.

Can I send my time domine data to you.

-- Thanks

[Thulasi Durai Durai Samy]

https://www.dropbox.com/s/iv1pbud4kzet30c/time%20domine%20data.tif

Time vs Acceleration response data

[Youssef Khmou]

hi, it looks like the frequency range is small, ok in this case you need to apply the Finite Fourier transform yo your data with 500 Hz as sample rate :

download the submission, apply the function and get the Bandwidth :

http://www.mathworks.com/matlabcentral/fileexchange/40002-psd-power-spectral-density-and-amplitude-spectrum-with-adjusted-fft

[Thulasi Durai Durai Samy]

Sir,

The code works fine, Can I have chat with you in gmail or skype, pls it will be so great full and sharing your great experience to beginners like me will help a lot

[Thulasi Durai Durai Samy]

Sir,

The code works fine, Can I have chat with you in gmail or skype, pls it will be so great full and sharing your great experience to beginners like me will help a lot

[Thulasi Durai Durai Samy]

Sir,

Actually I didnt make you clear, My aim is to reduce the noise in the time domine data and output I need in time domine data only. The noise is sitting on actual response signal as shown in figure. Pls let me know the solution pls, sampling rate: 500 Hz

href = ""<https://www.dropbox.com/s/iv1pbud4kzet30c/time%20domine%20data.tif ></a>

[Thulasi Durai Durai Samy]

https://www.dropbox.com/s/iv1pbud4kzet30c/time%20domine%20data.tif

The above hyperlink shows the graph