Re: Can I use Fractals for file/output scaling?

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



<x-charset iso-8859-1>Laurie,

The version of GF Print Pro that I downloaded was billed as 'fully
functioning' and it appears that it is.   The LE version is probably the one
that limits files to 40 Mb.

I took a small .jpg of a promotional drawing, and upsized it about 16x in GF
to about 43 MB, and did the same in Photoshop using bicubic resampling.
While neither could be described as 'good quality', the GF image has a
smoother look, and looks reasonable when viewed at a respectable distance.
The Photoshop one is unusable, in my opinion.  The PS file looks very
similar to the look of oversharpening, yet I didn't sharpen either one.  As
you said though, neither one enhanced the quality of the original image, GF
just produced a nicer looking result in this case.

There's a side-by-side at:
http://www.geocities.com/roberthaus/gftest.jpg

The GF one took over an hour on my 450 MHz machine, as opposed to about 10
seconds with Photoshop bicubic.

Gary Robertshaw
groberts@fix.net
Epson subscriber photo and art website list:
http://www.geocities.com/roberthaus/photosites.html


----- Original Message -----
From: "Laurie Solomon" <laurie@advancenet.net>
To: <epson-inkjet@leben.com>
Sent: Monday, June 12, 2000 10:19 AM
Subject: RE: Can I use Fractals for file/output scaling?


> >I'll use the remaining trials I have to test it out.
>
> Unfortunately, you cannot since the demo limits the size file and amount
of
> resizing that is possible.  You would have to get the full version to
handle
> original files sized over around 40MB and resizing of files over 4-5 X.
>
> I have stayed out of this discussion since the last time we had it there
was
> a definite decline in civility among the participants who for some reason
> seem to become emotionally invested in one product over another.  From
what
> I recalled of the discussion as to the quality of the resizing of GF
versus
> Photoshop, some enterprising individuals did some comparative test which
> they posted online.  Since the posted samples were all encoded into .jpg
> compressed formats, I questioned the validity of their evaluations (not of
> the evaluations of the originals before .jpg compression) as well as the
> fact that they would be viewed from low resolution monitors which could
have
> different effects on the way files displayed.  Questions concerning
> differences in the viewing distances used to examine a large printed image
> vs.. viewing a sample of that image in close up on a monitor arose as
well.
>
> The general conclusions were as you suggested.  Small amounts of resizing
of
> an image tended to show little difference between Photoshop's bicubic
method
> and GF methods, DEPENDING ON THE NATURE OF THE ORIGINAL IMAGE.  This
tended
> not to be as true as the amount of the resizing increased above 10X the
> original size, depending on the nature of the original image.  It was also
> noted that while GF's encoding and decoding process is very slow as
compared
> to other modes of compression and resizing, it did offer the convenience
of
> permitting near lossless compression of file sizes to smaller file sizes
> while also allowing for cropping and resizing of images from that
compressed
> file at the time of opening so that the same compressed files can be used
to
> produce a variety of final print sizes and croppings while maintaining the
> original compressed file.
>
> The was also a very cantankerous debate about whether or not GF resizing
> while maintaining or increasing/decreasing resolutions involved
> upsampling/downsampling in the same sense as the Bicubic method and the
> nature and quality of the effect of GF's process on the quality of the
final
> print (e.g., did it create artificial pixels as in the case of Bicubic
> methods when upsampling or did it work like a vector rendering program
that
> generated a entirely new rendering of the image at the new size based on
> mathematical formulas and not produce additional artificial pixels in the
> traditional sense like the other resampling methods).
>
> To conclude, I believe everyone needs to try the program out for
themselves
> and evaluate it in terms of the sorts of images that they are working with
> and the sorts of final results that they wish to produce.  However,
everyone
> should keep in mind that the one point of agreement among all the
> discussants and debaters was that GF WILL NOT improve image quality or
> enhance images; it will not correct or remedy poor quality images or
images
> whose original suffers from being too low in resolution, unsharp in
> character, or otherwise distorted.
>

>

-
* WIN either one of three Digital Cameras or Free Photoquality Prints!   *
*          http://www.pcphotoreview.com/go.cfm?ref=lebenpr               *
* Shop smart! Read product reviews of photo printers and digital cameras *

</x-charset>

[Index of Archives]     [Yosemite News]     [Yosemite Photos]     [Scanners]     [Gimp]     [Gimp Users]
  Powered by Linux