From: "Clemens Valens" <cvalens2@sodena.net>
Subject: Re: Fourier Transform - Sort Of
Date: Fri, 18 Aug 2000 16:01:04 +0200
Newsgroups: sci.math
Summary: [missing]

> Since the posting of the message you quote, I have been given a lead
> on transforms that selects frequency values as human hearing does:
> the Mellon Transform and the Fast Mellon Transform.  They have
> something to do with the wavelets transforms.  I think that they are
> particular implementations, but I have not found them yet.

This book has a long chapter on the Mellin (not Mellon) transform:

The transforms and applications handbook
Ed. by A. D. Poularikas, Boca Raton, Fl (USA): CRC Press, 1996.
The Electrical Engineering Handbook Series.

I am in the proces of (slowly) building a transforms web site (since it is
not yet very far I won't post the URL, but you can find it if you try) and
have yet to read up on the Mellin transform, so I may be mistaken, but the
Mellin transform was dicovered before the wavelet transform and as far as I
know they don't have anything special to do with eachother, except of course
that they both are transforms that relate the time and frequency domain in
some way.

Clemens
==============================================================================

From: Future Beacon <progress@eznet.net>
Subject: Re: Fourier Transform - Sort Of
Date: Fri, 18 Aug 2000 11:03:04 -0400
Newsgroups: sci.math

Clemens,

Thank you for this message and reference:


On Fri, 18 Aug 2000, Clemens Valens wrote:

> > Since the posting of the message you quote, I have been given a lead
> > on transforms that selects frequency values as human hearing does:
> > the Mellon Transform and the Fast Mellon Transform.  They have
> > something to do with the wavelets transforms.  I think that they are
> > particular implementations, but I have not found them yet.
> 
> This book has a long chapter on the Mellin (not Mellon) transform:
> 
> The transforms and applications handbook
> Ed. by A. D. Poularikas, Boca Raton, Fl (USA): CRC Press, 1996.
> The Electrical Engineering Handbook Series.
> 
> I am in the process of (slowly) building a transforms web site (since it is
> not yet very far I won't post the URL, but you can find it if you try) and
> have yet to read up on the Mellin transform, so I may be mistaken, but the
> Mellin transform was discovered before the wavelet transform and as far as I
> know they don't have anything special to do with each other, except of course
> that they both are transforms that relate the time and frequency domain in
> some way.
> 
> Clemens


I lost a lot of time by being given the wrong spelling.  Your
correction immediately produced results.  I have a bookmarks
file started for transformations at

http://eznet.net/~progress/dl/trans.htm

The last few entries lead to Mellin Transform references.

I now find out that the connection to wavelet transforms
is only that a professor thinks that there is a certain
similarity or common thread of reasoning.  I don't know if
that is true.

Does anybody know if there is an efficiency in visual compression
or processing analogous to that in auditory processing using the
Mellin Transform?

Thank you for your help.


Jim Trek
Future Beacon Technology
http://eznet.net/~progress
progress@eznet.net
==============================================================================

From: jan@panteltje.demon.nl (Jan Panteltje)
Subject: Re: wavelet transform
Date: Wed, 16 Aug 2000 19:32:28 GMT
Newsgroups: sci.math,sci.physics

There is a lot on the http://www.cs.kuleuven.ac.be/~wavelets/
Catholic university Leuven, in Belgium
There is also a PhD thesis there, by G Uytterhoeven 'wavelets software
and applications', and some real working C++ code.

About the lin - log conversion I thing mr. Lutus is right, it is just a
conversion before the display ?

I am interested in it for transmission of video data, since wavelet
compression is better then mpeg, with less artifacts.
Maybe something for the future TV system :-) hehe

Hardware solutions for video compression in real time in the form of
coder / decoder chips and demo boards:
http://www.analog.com/

Using a chip and then some software to do the display conversion to log
should not be impossible I think?
Would be really fast..
have not tried that though....
Jan