Yahoo Groups archive

Datacolor User to User Support Group.

Index last updated: 2026-04-28 23:18 UTC

Message

Re: "Saturation" versus "Perceptual" rendering intent -- problems with yellow

2008-07-25 by marko.mili

Yes. Please see http://www.cognistudio.com/profile_test/

OEM looks most natural, at the cost of blown highlights. Spyder looks bad color-wise. I 
didn't manipulate scanned images in any way other then assigning them scanner icc 
profile, but I would say that prints are even lighter then shown here, so OEM print is way 
too light, while Spyder prints match original digital photo in luminosity.

Marko

--- In colorvision_group@yahoogroups.com, Cdtobie <CDTobie@...> wrote:
>
> Have you turned off BPC?
> 
> C. D. Tobie
> WW Product Technology Mngr.
> Digital Imaging & Home Theater
> DataColor.com
> CDTobie@...
> 
> On Jul 24, 2008, at 4:52 PM, "marko.mili" <marko.mili@...> wrote:
> 
> > This is an old thread, but my question ties into it. I have been
> > trying to profile Canon iP4500 and iP4300 printers, with both OEM and
> > Image Specialists inks. It quickly becomes obvious that these printers
> > suffer in reproduction of dark saturated colors, especially reds and
> > to lesser degree greens. Not so bad in yellows and blues. With OEM
> > profiles, the way Canon gets around this is by having perceptual
> > rendering quite a bit lighter then actual colors, which brings dark
> > colors into gamut. This is just eyeballing soft proofs and a few
> > output samples. I don't know really if Canon does this with images
> > when all colors are in gamut. These printers suck and most images have
> > plenty out of gamut. The main difference is in how OEM and Colorvision
> > differ in perceptual intent, with OEM being about a stop too light.
> > Other intents look about the same between OEM and Colorvision.
> >
> > The end result of this, due to how poorly these printers render dark
> > saturated reds and browns is that none of intents made by Colorvision
> > look photographically natural. Not that colors don't match source, but
> > relationships between colors look unnatural. For example, if someone
> > is wearing a red sweater, in folds of the fabric where red is in a
> > shadow, it looks weirdly brown/grayish in all intents (some more some
> > less, but none satisfactory). With OEM perceptual, all colors are
> > wrong (mainly in that they are too light, but also in hue and
> > saturation), but at least they preserve relationship between colors
> > and it looks believable.
> >
> > I would suggest that Colorvision perceptual rendering shouldn't match
> > on just hue and brightness and ignore saturation (if that is how it
> > works, although I suspect that explanation given below is simplified).
> >
> > In a difficult situation at least the way OEM perceptual works is
> > effective. How can similar results be achieved with Colorvision
> > profiles, without bringing each image into gamut by hand (which is
> > fine for exhibition, but not when I want to print 100 vacation
> > snapshots for my mom)?
> >
> > Marko
> >
> > --- In colorvision_group@yahoogroups.com, CDTobie@ wrote:
> >>
> >>
> >> In a message dated 11/14/07 12:31:47 PM, ryannd@ writes:
> >>
> >>
> >>> On the profile I made with PrintFixPro the bright yellows
> >>> (especially around the sun in the sunset picture at the lower left
> >>> but all yellows in other parts too) are muddy and have lots of black
> >>> in them on perceptual and look good on saturation intent. And the
> >>> prints look just like the soft proof--unusably muddy on the yellow
> >>> colors with perceptual rendering.
> >>>
> >>> You guys off course recomend saturation and not perceptual intent on
> >>> rendering and that works fine for CS3.  Unfortunately I do almost
> >>> all my printing with Lightroom which doesn't give me a choice of
> >>> saturation -- it's perceptual or relative only as rendering intent
> >>> with Lightroom.
> >>>
> >>> Help!
> >>>
> >>
> >> In one sense, this is a Lightroom issue, in that they do not support
> > the full
> >> list of intents. What is happening is that you are asking for out of
> > gamut
> >> yellows, and the match is being selected based by Hue and
> > Brightness, without an
> >> emphasis on Saturation, in our Perceptual intent. The color with the
> > closest
> >> density and hue angle does not suit you. Between now and the next
> > major update
> >> to Spyder3Print, there will not be any changes to our color engine,
> > so what
> >> you get is what you get, for the time being. This is only one of
> > several
> >> reasons I don't print from Lightroom.
> >>
> >> On the other hand, the test images you note are filled with
> > oversaturated
> >> colors, colors that you don't actually find in photos, only in
> > vector images and
> >> heavily adjusted images. This test image has lots of yellows that
> > hit 255 red,
> >> and well over 200 green in AdobeRGB. Not colors you're likely to
> > bump into
> >> with actual images. So be sure this problem is a problem for your
> > own files
> >> before you panic, based on this over-the-top test image. Don't judge
> > the
> >> "reasonableness" of these test image yellows by looking at them on
> > your monitor. They
> >> are all outside the gamut of a typical display; most are also
> > outside the gamut
> >> of a Wide Gamut Eizo display...
> >
> >
> >
> > ------------------------------------
> >
> > Yahoo! Groups Links
> >
> >
> >
>

Attachments

Move to quarantaine

This moves the raw source file on disk only. The archive index is not changed automatically, so you still need to run a manual refresh afterward.