Fourier Shift theorem python

Viewed 108

So I am trying to use the shift theorem to to find the Fourier transform of two square waves that are separated by 30 pixels, but when I compare the Fourier transform of the two waves to the Fourier transform of one wave added to second wave using the shift theorem they don't match, and I have no idea why

import numpy as np
import matplotlib as mpl
import matplotlib.pyplot as plt

plt.rcParams['image.cmap'] = 'magma'

import h5py 
import scipy.fftpack

%matplotlib notebook
def magsq(arr):
    return np.abs(arr*np.conj(arr))

fig1,ax1=plt.subplots()
l=512
test=np.zeros(l)
a=int(len(test)/2)
b=10
test[a-b:a+b]=1
test
ax1.plot(test)

test2=np.zeros(l)
test2[a-b:a+b]=1
test2[a+b+30:a+b+30+2*b]=1
ax1.plot(test2,'--')

FT=np.fft.fft(test)

x=np.linspace(0,l,l)
w = np.fft.fftfreq(30)
s=30
shift=FT*np.e**(1j*2*np.pi*x*s/l)

shifted=FT+shift
FT=np.fft.fftshift(FT)
shifted=np.fft.fftshift(shifted)
shiftedi=magsq(shifted)

FT2=np.fft.fft(test2)
FT2=np.fft.fftshift(FT2)
FT2I=magsq(FT2)
fig2,ax2=plt.subplots()
#ax2.plot(FTI)
ax2.plot(shiftedi)
ax2.plot(FT2I,'--')

enter image description here

1 Answers

Your code to shift the signal by 30 pixels should be as follows.

test2[a-b+30:a+b+30]=1

Then both transforms exactly match

enter image description here

Related