Re: Rendering intents: Perceptual vs. RelCol
Re: Rendering intents: Perceptual vs. RelCol
- Subject: Re: Rendering intents: Perceptual vs. RelCol
- From: email@hidden
- Date: Fri, 26 Jan 2001 12:49:57 EST
In a message dated 1/26/01 9:59:22 AM, email@hidden writes:
>
For the sake of argument, let's assume I'm using Adobe98 as my RGB WS and
>
converting to some "PressCMYK" WS that is, of course, quite a bit smaller
>
than my RGB WS.
>
>
Let's also assume I have two RGB images, one that is very "out-of-gamut"
>
relative to my PressCMYK WS ("Image A") and one that has colors that are
>
totally in-gamut relative to my PressCMYK WS ("Image B").
>
>
What happens to "Image B", my totally in-gamut image:
>
>
RelCol: Nothing much happens here. Since all colors are in-gamut, everything
>
gets mapped one-to-one with no clipping and overlapping of colors.
Correct
>
>
(the question:)
>
Perceptual: What happens here? Does A) all the colors *still* get compressed
>
because there's no mechanism for it deduce that all the colors are in-gamut
>
so no compression is necessary or B) all the colors get compressed anyway
>
because it has to *assume* that there are colors stretching out to the
>
edges
>
of the RGB WS? In other words, if you have RGB WS "A" and CMYK WS "B" and
>
CMM "C", every image, regardless of content or gamut, gets "compressed"
>
exactly the same way?
Exactly, Perceptual gamut mapping and compression are not image specific.
That would require a very smart CMM and a lot of processing time to check
each image individually. The result would be less compression on ow gamut
images, but letsl look a a specific case: a foggy landscape shot, then the
same shot with all settings identical, but a hunter in a blaze ornage vewst
has walked into view. Now if you print hte images side by side in a magazine,
they will not look similar, sa they would have beofe, but instead the one
with the hunter in it will have less saturation in all the otehr portions of
hte image. I don't think we are quite ready to have these decisions made
for(the "samrt CM" alternative). but need to decide on a case by case basis
as needed.
>
I've always assumed that Perceptual was preferable
>
but
>
I'm now wondering if RelCol isn't a better default (more "accurate"?) and
>
from there just use the Gamut Warning as way to decide whether Perceptual
>
should be considered. So RelCol is the default and Perceptual is used on
>
an
>
as needed basis. Make sense or am I nuts?
>
If you are willing to make a few reasonable decisions about images, or sets
of images, based on image content, and to use the out of gamut warning and
profile based soft proofing in PS6, then you will indeed be able to improve
your results over mindless Perceptual conversion.
C. David Tobie
Design Cooperative
email@hidden