I came across literature which says that the Fourier - Bessel (Hankel) transformation is better than a STFT or a Wigner - Ville distribution as the Bessel function forms an orthonormal basis and they occupy finite bandwidth. This is supposed to translate into better time and frequency resolutions. Isn't it the same with the regular fourier transform ? so why not use an STFT itself to transform the signal into time frequency domain ?

  • $\begingroup$ Could you provide information on where you came across such claims? $\endgroup$ – Laurent Duval Jan 25 '18 at 7:31
  • $\begingroup$ Hey Laurent, I may be wrong as well. I am just starting off, and want to understand better: pdfs.semanticscholar.org/468a/… $\endgroup$ – Asp Jan 25 '18 at 13:28

The Hankel Transform includes the Bessel function as part of its definition. I wouldn’t say the Bessel has an advantage. I would say that the Hankel Transform has an advantage where a two dimensional signal has radial symmetry, or saying it another way, the signal in $x$ and $y$ can be expressed as a product of two functions in $r$ and $\theta$ through a change of variables. $x=r\cos \theta$ $y=r\sin \theta$

$$f(x,y)dx dy \Rightarrow g(r) h( \theta) r dr d \theta $$

The Hankel Transform is equivalent to a Fourier Transform on the radius $r$ variable.

Any advantage over STFT or Wigner would need to be explained in the literature you read. Im not aware of any general use of Hankel grams, or why the Hankel would make any obvious relevance outside of two dimensional geometries with rotational symmetry.

| improve this answer | |

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.