On 21-Jul-2003, Joao S. O. Bueno wrote:
> On Monday 21 July 2003 4:47 pm, Adam D. Moss wrote:
> > Patrick McFarland wrote:
> > > So, if gegl isnt going to be in gimp2, when will it be?
> > >
> > > Ive been waiting for gimp2 awhile now, and now that gegl wont be
> > > in it, I have to keep waiting. How long will I have to wait now?
> > > 2.2? 2.4?
> > gegl isn't a panacea...
> Perfectly said.
> Actually, I've skimmed trhough some docs on the GEGL, and I wonder,
> what are its actual uses for the final user?
> I can see it provides the grounds for easier hacking in the GIMP, and
> will facilitate the implemanetation of internal CMYK and FLoating
> point images, and such. But for GEGL alone, what does the artist
Ive been having a long discussion with Robin Rowe about stuff he possibly
wants to do with CinePaint. I am, in this context, an artist.
Stuff like CMYK is obvious how thats useful. (Desktop -> printer stuff, etc)
Floating point, however, is harder to wrap your mind around. The short answer
is that quantiziation of data is bad, and more precision is better.
To continue on this line of thinking, integers are limited to 0.0 and 1.0
(In 8bit per channel images, this is 0 through 255). Once its goes above 1.0,
or below 0.0, data is lost. Floating point values are not clipped, so data is
In addition, gegl would have been able to internally calculate using higher
precision values, which allows it to unresrect itself from the "accumulating
error." (Basic programming concept, google for it.) So, you may have 8 bits
per channel, but over the course of a dozen layers, you have tons of data.
(More than can be represented in 8 bits at all times.)
To avoid the accumulating error from snowballing, you use a much higher
precision to calculate it. Floating point math comes in handy here too.
So, with floating point, we've avoided clipping, and we've avoided premature
quantization and accumulating of errors.
Of course, most of what I said above applies to what you call "layer
combining." To apply floating point goodness to individual layers, take the
common arguments for 8 bit per channel vs 16 bit per channel, and replace 8 bit
with the word "integer", and 16 bit with the phrase "floating point". _In
addition_ to the higher precision (called a larger dynamic range in film talk),
you also can ignore clipping, because floating point doesnt have it.
As an artist, the high precision layer combining is a godsend, and as a very
gimp-abusive artist, high precision and no clipping all around is a godsend.
So, this is why Im disapointed that it never came into being.
Patrick "Diablo-D3" McFarland || [EMAIL PROTECTED]
"Computer games don't affect kids; I mean if Pac-Man affected us as kids, we'd
all be running around in darkened rooms, munching magic pills and listening to
repetitive electronic music." -- Kristian Wilson, Nintendo, Inc, 1989
Gimp-developer mailing list