MATLAB: Instead of using the threshold to select coefficients in the Fourier domain, you could also use a low-pass filter. can some one explain which would be better

low-pass filter

Instead of using the threshold to select coefficients in the Fourier domain, you could also use a low-pass filter. can some one explain which would be better?

Best Answer

I think using a lowpass filter is better in most cases than just zeroing out coefficients in the Fourier transform.

A lowpass filter gives you a better attenuation profile by multiplying the signal's Fourier transform with the Fourier transform of the filter (convolution in time).

Related Solutions

MATLAB: Low-pass filter

hsize=??          % define
h=fspecial('average',hsize);
filter_wmarked=imfilter(wmarked,h);

MATLAB: Ideal low-pass filter transfer function

The ‘ideal’ filter is symmetrical about the origin in the frequency domain (after doing the fftshift call), so will be 1 from -25 Hz to +25 Hz, and zero elsewhere. Define it as such, and then do the multiplication with the signal in the frequency domain. Then use ifftshift on the filtered signal, then ifft to return it to the time domain.

Be sure to scale the result correctly, since filtering removes much of the energy in the original signal.

Best Answer

Related Solutions

MATLAB: Low-pass filter

MATLAB: Ideal low-pass filter transfer function

Related Question