Talk:Wavelet transform

Latest comment: 2 years ago by LutzL in topic Completeness of a basis

Wavelet transform

edit

I like the information contained in the article, but I find that Wavelet transform would be a more appropriate title for it. At least from the perspective of the signal analysts, but also makes sense from a very mathematical point of view, as it is a more clear concept. —Preceding unsigned comment added by 161.116.80.71 (talkcontribs) 15:48, 22 September 2006

At Talk:wavelet I made some suggestions about restructuring the whole tree of wavelet articels. So far there were no comments and also no time on my side to implement those changes. Wavelet transform is or should be an article that describes the distinction between continuous wavelet transforms and discrete wavelet transforms.--LutzL 15:54, 22 September 2006 (UTC)Reply

P-adic wavelets

edit

Maybe there should be a link to P-adic wavelets, basis for L^2(Q). It would be nice with a comparison of P-adic wavelets and ordinary wavelets.

Keep this separate

edit

It is good to keep these articles separate. The article on the wavelet series can show various series derivations and the motivations without cluttering the wavelet transform article with this math. The transform article would go on to indicate useful transform properties and utility thereof.

The Fourier series of articles (pun intended) is a good example set. Poppafuze (talk) 19:11, 6 June 2008 (UTC)Reply

Scope?

edit

To anyone still interested in this article, esp. the movers: What is the intended scope of this article? At this point, it has a section hinting at the discrete wavelet transform followed by a section about the continuous transform, both very incomplete and without any connection.--LutzL (talk) 09:26, 19 July 2010 (UTC)Reply

Old page history

edit

Some old page history that used to be at the title "Wavelet transform" can now be found at Talk:Wavelet transform/Old history. The corresponding talk page for this history can be found at Talk:Wavelet transform/Old talk page. Graham87 03:57, 14 May 2011 (UTC)Reply

The old wavelet compression hoax.

edit

Back in the early 90's when wavelet compression was first thought up, there was a company that wrote some compression software with claims of extreme lossless compression. It turned out to be a hoax. http://www.cv.nrao.edu/fits/traffic/compcompression/faq_1.news Look for The WEB 16:1 compressor in that document. Bizzybody (talk) 04:29, 5 March 2012 (UTC)Reply

Now please tell us if there is some recent history of this hoax. And why do you invoke wavelet compression, the FAQ for this item [9] does not contain a hint of wavelets.--LutzL (talk) 12:19, 5 March 2012 (UTC)Reply

Graphic comparing STFT and WT is wrong

edit

The grafic comparing STFT and WT is wrong. Because when scaling the wavelet it does not only cover a wider timespan but also its frequencyspan is reduced. — Preceding unsigned comment added by JuanchoArroyo (talkcontribs) 12:21, 11 September 2013 (UTC)Reply

I think that this is the case of discrete wavelet transform, not the case of wavelet transform in general (about which the article is). --DaBler (talk) 12:55, 6 November 2019 (UTC)Reply

"Discretization?"

edit

The "Other practical applications" section refers to "discretizing" a signal. I think the author meant "quantizing" and "quantization." Is it reasonable to make that substitution? — Preceding unsigned comment added by 70.88.250.17 (talk) 16:20, 16 June 2016 (UTC)Reply

Completeness of a basis

edit

The article seems to imply a wavelet basis is complete if any function can be represented exactly in the basis. What is meant by “basis” if one can be something other than complete? 018 (talk) 01:50, 29 September 2021 (UTC)Reply

The article is still a mess. The first section handles the special case of orthogonal 2-channel wavelets without denoting it as special case. It is also very sparse on the "transform" aspect. If your comment is on that section, then it may be a misunderstanding, the paragraphs on orthogonality and completeness explain the conditions to call the affine frame generated by   a Hilbert basis.
The next section "Principle" is on general sampling theory (of the continuous wavelet transform  ) without calling it so. The remaining sections are a collection of random highlights of other interesting facts and applications that are only loosely connected to the main topic "wavelet transform".--LutzL (talk) 08:21, 5 July 2022 (UTC)Reply