Windowing a signal in frequency space

mdornfe1

I'm trying to write a python script that takes a noisy multi frequency signal, transforms it to frequency space, windows it there with a gaussian, then transforms it back to time space. Here is what I wrote:

Code:

Fs=1000     #sampling frequency
fo=120      #center of gaussian   
sigma=0.01  #inverse width of gaussian
T=1./Fs
L=2**10     #number of samples
t=arange(0,L)*T #time vector
f=Fs*linspace(0,1,L)    #frequency vector
x=0.7*sin(2*pi*50*t) + sin(2*pi*120*t)+randn(t.size)/sqrt(t.size)   #signal
x_fft=fft(x)
W=exp(-square(2*pi*sigma*(f-fo)))   #gaussian window
y=ifft(W*x_fft)                      #windowed signal

The problem I'm running into is the windowed signal y has non negligible imaginary parts. They're about the same order as the real parts. Does anyone know why I might be getting this?

Bill Simpson

Any chance you are getting a complex vector when you were expecting a,b as in a*cos(n*t)+b*sin(n*t)?

Or are you perhaps thinking you want the magnitude at each frequency and not the phase information?

mdornfe1

I just realized what was happening. When I multiplied the transformed signal by the gaussian, I cancelled out the negative frequency components. So of course when I transformed back the signal would be complex. I was expecting it to be a real sinusoid not a complex one.

Similar discussions for: Windowing a signal in frequency space
Thread	Forum	Replies
Signal Windowing Help	Engineering, Comp Sci, & Technology Homework	0
how to convert time domaine signal into frequency domain signal	Electrical Engineering	3
signal frequency!!!	General Engineering	5
frequency spectrum of the modulated signal g(t)	Engineering, Comp Sci, & Technology Homework	3
Low frequency signal	Electrical Engineering	2

Bill Simpson	Any chance you are getting a complex vector when you were expecting a,b as in acos(nt)+bsin(nt)? Or are you perhaps thinking you want the magnitude at each frequency and not the phase information?

mdornfe1	I just realized what was happening. When I multiplied the transformed signal by the gaussian, I cancelled out the negative frequency components. So of course when I transformed back the signal would be complex. I was expecting it to be a real sinusoid not a complex one.