I need to study a wave file for every 100 Hz. but I do not know how I can design a filter for 100 Hz intervals (0-100 Hz, 100-200 Hz, ... , ((fs/2)-100) _ (fs/2) Hz).
Is there MATLAB code for this? How many samples are needed for a bandwidth of 100 Hz? How should calculate it? Thanks in advance.
It is directly depends on the sampling frequency to decide how many samples use and also the interval you want to filter is also important. for example it is different for interval 200-300 Hz or 2000-2100 Hz.
one way to filter a signal is to take FFT in Matlab and select the interval you want.
other way is to use fdatool in MATLAB and design your favorite filter.
The 'bandpass' range of the FFT-based filter has the shape of a sinc function but the FFT will, of course give you a first idea of the power spectral density. If you collect the signal over a period To then the width of the main lobe of the sinc function is inversely proportional to To. If your signal was sampled over a time window of length To (value 1 from -To/2 to +To/2) then the Fourier frequency estimator is a sinc function of the form X(f)=sin(pi*f*To)/(pi*f) (which is the Fourier transform of the rectangular window. Of course if a bandpass filter with the gain with the shape of a sinc function is not satisfactory and you need a better comb filter, with the gain more similar to a rectangular shape (gain approximately unity over the passband) then you have to design the individual bandpass filters. I hope the above makes sense to you and is helpful. Best wishes.
- Badges
- Science topic