0
$\begingroup$

According to Fourier transform, a continuous wavelet could be written as $$ \psi(t)=\frac{1}{2\pi}\int\hat\psi(k)\text{e}^{-ikt}\text{d}k $$

From the equation above, we know that $\psi(t=0)$ is $$ \psi(0)=\frac{1}{2\pi}\int\hat\psi(k)\text{d}k $$

Is it possible that $\psi(0)=0$?

$\endgroup$
1
$\begingroup$

Yes: a sufficient criterion for a valid wavelet is being zero mean in time domain, i.e. $\hat \psi (0) = 0$. For $\psi(0) = 0$ we require that $\hat \psi$ sums to zero; one example is higher order Generalized Morse Wavelets - from Olhede & Walden,

(Though above is only approximately zero-sum). Such wavelets can be admissible and analytic, enabling CWT inversion. Trivial Python example below; GMWs are also implemented in ssqueezepy.

enter image description here

Code

import numpy as np
import matplotlib.pyplot as plt
from numpy.fft import ifft, ifftshift

t = np.linspace(0, 1, 256, endpoint=False)
wf = np.exp(-(t - .08)**2 * 4096)
wf -= np.roll(wf, 25)    # shift by 25 samples and subtract
w = ifftshift(ifft(wf))  # take to time domain and center

plt.plot(wf)
plt.show()

plt.plot(w.real)
plt.plot(w.imag)
plt.show()
$\endgroup$
0
$\begingroup$

Yes, perfectly possible. That means that the mean of the function $\hat{\psi}(k)$ is zero.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.