Re: smooth gradations
Re: smooth gradations
- Subject: Re: smooth gradations
- From: neil snape <email@hidden>
- Date: Thu, 06 Dec 2001 10:19:34 +0100
on 05/12/2001 09:06, joseph wilhelm at email@hidden wrote:
>
> While I've had surprisingly good results with producing my own custom
>
> profiles for use with my Best rip and Epson 7500, my output still seems to
>
> suffer when faced with a gradient. Seemingly smooth gradients print with
>
> banding and what should be a gradual fade turns into an abrupt ending.
>
>
>
I have this same problem. I print landscape photographs and clear
>
blue skies will show banding. The affect is exaggerated when I
>
transform from my working space to a custom printer space - even
>
without color correction. I use an Epson 7500 and its OEM driver but
>
I don't think it is printer specific. I have noticed that different
>
CMMs reveal banding differently and some reveal it less. In my case
>
the banding can be seen on screen so it is easy (although quite
>
subtle) to preview.
>
>
I have been trying to solve this problem too and recently it has been
>
suggested
>
that I ought to work in 48 bits. Another idea may be to flatten out
>
the curve some in the original scan in the area that is banding.
These are two different problems resulting in a similar appearance. Sounds
that that metamerism debate again!
Banding by the printer can be and often is the firmware plotting points with
abrupt differences , such as blue skies when the cross over of light c
becomes C.
If the scans have banding already then your scanner (mine too) is not seeing
even step differences. With Heidelberg CCD scanners you can smooth the light
tones with base density. This can smooth some differences in the light areas
of scans. Normally a drum by nature of PMT won't band because of step
variations. There has been big improvements in the profile smarts that have
mostly solved early color hue shifting to purple, but this is an ongoing
dilemma as it's CMM dependant and color model dependant.
Maybe someone could tell us if with the Colorspan ? with independent start
points for the individual channels can eliminate banding by changing the
settings here.
Neil Snape email@hidden
http://mapage.noos.fr/nsnape