squared fast Fourier transform of a moving boxcar filter

I need to calculate time evolving PSD or Squared fft of a signal at certain frequency using a moving boxcar filter. I need to show the variations of signal at certain frequencies over time.

Answer 1

[Wayne King]

Just specify the boxcar filter as the window argument in spectrogram.

t=0:0.001:2;                   
x=chirp(t,0,1,150);  
win = 1/200*ones(200,1);
% if you want unit norm in your filter
win1 = 1/sqrt(200)*ones(200,1);
[y,f,t,p] = spectrogram(x,win,100,length(win),1E3);

The matrix, p, contains the short-time PSD estimates for a length 200 with a 100 sample overlap in the above example. Tailor it as appropriate for your case.