Re: TOTALLY OFF TOPIC: 24-bit-per-pixel limit to Gdk RGB ?

[TORRI Vincent <Vincent Torri math u-bordeaux fr>]

On Wed, 19 Feb 2003, individual wrote:

> On Wednesday, February 19, 2003, at 12:06 PM, TORRI Vincent wrote:
> > > > http://www.yorku.ca/eye/machband.htm
> >
> > One thing about this link : the fact presented at this link looks like 
> > the
> > Gibbs phenomenum in the Fourier analysis. Quite interesting, as we can
> > consider that the eye make a Fourier transform, according to some 
> > papers
> > (the eye detects the wavelength (or the frequencies, as you want) of 
> > the
> > colors)
>
> I seriously doubt it. The Gibbs "overshoot" appears in the frequency 
> domain, not in the spatial domain, which the lower graph represents.

Well ? I don't think so : if you take one 1-periodic funtion which is NOT 
continuous (called f), f(k) its kth Fourier coefficient and Fn(x) the sum 
over k=-n to n, of f(k)exp(2iPikx), and if you display Fn, you will see 
the Gibbs phenomenum (L infty norm of f < a constant < L infty norm of 
Fn).
  Take f(x)=0 for x in [0;0.5[ and f(x)=1 for x=0.5;1] (which is almost 
the function drawn in the page below) and draw Fn (n=50 for example). I 
will look (a bit of course) like that.If you take f as the black function, 
and do the same computations, you draw Fn and you remark that Fn goes 
upper and lower than f (even if there is no Gibbs phenomenum in this 
case).
 The Gibbs phenomenum appears in the "spatial" domain.
  My previous post was just a remark : the red curve let me think about 
the gibbs phenomenum. I don't know any link between Gibbs phenomenum and 
the mach bands.

regards

Vincent TORRI


> Paul
>
> _______________________________________________
> gtk-app-devel-list mailing list
> gtk-app-devel-list gnome org
> http://mail.gnome.org/mailman/listinfo/gtk-app-devel-list

-- 
TORRI Vincent
Mathematiques Appliquees Bordeaux                                                
Institut de Mathematiques                                                       
Universite Bordeaux 1                                                           
351 cours de la liberation                                                      
33 405 Talence cedex - France                                                   
                                                                                
Tel : 33 (0)5 57 96 21 42                                                       
Fax : 33 (0)5 56 84 26 26                                                       
--