Pages: 1 [2] 3   Go Down

Author Topic: Why are dark blues purplish in test print  (Read 9722 times)

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Why are dark blues purplish in test print
« Reply #20 on: January 11, 2018, 10:15:59 pm »

Mark,

One thing I've used myself and recommend to others is to grab a ColorChecker image file from Babelcolor and print it using Abs. Col. intent. It can be compared to an actual Colorchecker which nearly everyone has. This allows checking a known (Colorchecker) to a printer/paper/profile and is a good way to check against gross errors on the printer side.  Best to use a low OBA paper but otherwise is a really good check.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Why are dark blues purplish in test print
« Reply #21 on: January 12, 2018, 09:38:31 am »

Doug,

Agreed; as you know, I've been doing this formally (i.e. statistically) in umpteen paper reviews I've conducted for this website going back quite a while now; but your suggestion of a visual check is fine for those that can't be troubled with spreadsheets and formulas. That said, it has a potential limitation on which I'm now starting fresh work - and that is the gamut question. As you know, the 24 patch CC is a "safe" rather narrow gamut target - likely designed that way intentionally many years ago when the gamut volume of devices and processes were generally less than they are today, in order to avoid confounding accuracy testing with the handling of OOG colours; that's one important reason why it's so widely used for these purposes. So, in this particular case, if the described blue to purple colour shift were happening close to the gamut boundary of the printer/paper profile because the profile is coming short on accurate rendition of very saturated blues, the CC24 check wouldn't necessarily reveal this. To illustrate what I mean, see the attached, where the the CC24 is mapped against the IGFS/P800 gamut volume, and the red arrow shows where the "Sky Blue patch" sits within that volume - well ensconced, such if the O/P's problematic blues were closer to the gamut boundary, the CC24 wouldn't be revealing the potential source of the problem, e.g. distortion of highly saturated blue.

Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Why are dark blues purplish in test print
« Reply #22 on: January 12, 2018, 10:42:42 am »

You can print Absolute Colorimetric but if you wish to use say Perceptual and you witness a magenta shift, I suspect you've got a goofy profile.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Why are dark blues purplish in test print
« Reply #23 on: January 12, 2018, 01:23:12 pm »

Doug,

Agreed; as you know, I've been doing this formally (i.e. statistically) in umpteen paper reviews I've conducted for this website going back quite a while now; but your suggestion of a visual check is fine for those that can't be troubled with spreadsheets and formulas.
Your reviews are quite informative and technically detailed but we are fortunate to have the instruments and technical background to do this. The OP, and perhaps the large majority of folks that have purchased high quality printers only have their eyes. No matter how good a person is with spreadsheets and formulas none of that helps when they look at a print and the color seems off.  There are a few areas where color perception is both quite picky, and pretty precise. Skin colors and other memory colors. And there are areas where hue shifts are more noticeable such as the blue-purple area.

Badly needed are physical, and accurately reproduced, reference images printed using Rel. Col. that can be purchased or distributed as samples with new printers, combined with the downloadable digital images. The image should be in gamut for a wide selection of printers for a given media type. It's important that they are in gamut because out of gamut mapping is completely uncontrolled. Fortunately, profiles that are accurate over a large set of within gamut colors will also show accurately (soft proofing and/or math) how they render out of gamut colors as the AtoB1 tables are the quite accurate*. They can then be evaluated for how pleasing (or not) those out of gamut color mappings appear. OOG rendering is a different issue that verifying the printer/profile accuracy.

A Colorchecker card is the closest thing we have to a reference image with controlled colors that almost all photographers possess until something like the above is available.

* Aside from profiles created where the RelCol tables incorporate BPC like some Epson canned ones, unfortunately. However, while this screws up soft proofing, amongst other things, it has minimal effect on colors with L* values materially larger than the paper's black point.
Quote
That said, it has a potential limitation on which I'm now starting fresh work - and that is the gamut question. As you know, the 24 patch CC is a "safe" rather narrow gamut target - likely designed that way intentionally many years ago when the gamut volume of devices and processes were generally less than they are today, in order to avoid confounding accuracy testing with the handling of OOG colours; that's one important reason why it's so widely used for these purposes. So, in this particular case, if the described blue to purple colour shift were happening close to the gamut boundary of the printer/paper profile because the profile is coming short on accurate rendition of very saturated blues, the CC24 check wouldn't necessarily reveal this.
A working profile will reasonably accurately tell you exactly what colors will be printed from any of the rendering intents regardless of how near or far from the gamut boundary the color is. The AtoB1 tables are well defined and very accurate. At least for those that follow the last V2 ICC spec. or later. :)
Quote
To illustrate what I mean, see the attached, where the the CC24 is mapped against the IGFS/P800 gamut volume, and the red arrow shows where the "Sky Blue patch" sits within that volume - well ensconced, such if the O/P's problematic blues were closer to the gamut boundary, the CC24 wouldn't be revealing the potential source of the problem, e.g. distortion of highly saturated blue.

Yes, it would be nice to have an alternative for people without spectrophotometers. Something well defined and controlled with more than 18 colored patches.
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Why are dark blues purplish in test print
« Reply #24 on: January 12, 2018, 01:28:07 pm »

You can print Absolute Colorimetric but if you wish to use say Perceptual and you witness a magenta shift, I suspect you've got a goofy profile.
Which is why one needs to stick with Rel. Col. or Abs. Col. and in gamut images. "Accurate" only applies to those. If it's accurate there, but goofy in Perceptual, time to get a profile from someone else. You can't fix goofy when it's baked into the profile.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Why are dark blues purplish in test print
« Reply #25 on: January 12, 2018, 01:42:09 pm »

..............A working profile will reasonably accurately tell you exactly what colors will be printed from any of the rendering intents regardless of how near or far from the gamut boundary the color is. ................

There is an argument (from a number of expert sources) that the 18 colours in the CC24 are very useful but "low hanging fruit"; so I am taking your statement as the proposition to be tested rather than a confirmed statement of fact. If the tests show it to be correct, that would confirm your statement at least in my mind and make me happy.

The relevance of this issue to the O/Ps problem of course relates to the unconfirmed proposition that hue shifts may be a function of high colour saturation that the profile is not handling well either because the blue is OOG and not being appropriately compressed into the printable gamut, or the blue is say barely in-gamut which the profile is also not handling well. Or what Andrew more directly calls a "goofy" profile. Lending credibility to the profile being the problem is that Printer Manages Color renders the blues more accurately according to the visual perception of the O/P, but distracting from the profile being the problem is the O/Ps observation that another profile has reproduced the problem, so two goofy profiles? Maybe, but maybe not. This is a tough nut to crack, so far.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Why are dark blues purplish in test print
« Reply #26 on: January 12, 2018, 02:24:42 pm »

There is an argument (from a number of expert sources) that the 18 colours in the CC24 are very useful but "low hanging fruit"; so I am taking your statement as the proposition to be tested rather than a confirmed statement of fact. If the tests show it to be correct, that would confirm your statement at least in my mind and make me happy.

The relevance of this issue to the O/Ps problem of course relates to the unconfirmed proposition that hue shifts may be a function of high colour saturation that the profile is not handling well either because the blue is OOG and not being appropriately compressed into the printable gamut, or the blue is say barely in-gamut which the profile is also not handling well. Or what Andrew more directly calls a "goofy" profile. Lending credibility to the profile being the problem is that Printer Manages Color renders the blues more accurately according to the visual perception of the O/P, but distracting from the profile being the problem is the O/Ps observation that another profile has reproduced the problem, so two goofy profiles? Maybe, but maybe not. This is a tough nut to crack, so far.

For what it's worth, I just made two prints of Andrew's image on Epson Prem. Luster 250 gsm using Rel Col on my 10+ y/o 9800. One used the canned Epson profile for that paper and one using a new profile made using my Isis. There is no visible difference outside of a very slight difference in rendering of the rainbow gradients used at the top and in some of the lettering. Big difference when viewing soft proofs but even there, both soft proofs are reasonably close. The woman, colorchecker, and image with the hand/jewelry looks the same on both prints.

I then ran a Matlab script on all the image pixels and compared the Lab values reported by the device space image RGB pixels comparing the canned profile and my custom profile.

Average dE00 1.2
Average dE00 w/o the OOG colors, 1.0.
Max dE00 (including OOG rainbow portions) 6.0

Average dE76 1.8
Average dE76 w/o the OOG colors, 1.5.
Max dE76 (including OOG rainbow portions) 8.1

Note that the test image doesn't have significant amounts of very low L* areas which are where the defects in Epson's canned profiles show up "bigly" (see Andrew's stress test image with Bill's Balls).

I've always found Epson's canned 9800 (and earlier 2400) profiles to be quite good but I didn't expect this close a match. Typically enough variation just between spectro brands to produce errors in this range and this was a 10 y/o profile and printer. I've only used OEM inks but haven't been particularly kind to my printer otherwise.


Note to the OP: If you PM me a mailing address I can mail you a print so you can physically compare with what you are getting.

EtoA: The blue jewelry shows no sign of purple/violet and, as noted, looks the same in both prints. The hue is really a long, long way from anything purple. It's closer to a cyanish blue if anything.
« Last Edit: January 12, 2018, 02:37:23 pm by Doug Gray »
Logged

Stephen Ray

  • Full Member
  • ***
  • Offline Offline
  • Posts: 217
Re: Why are dark blues purplish in test print
« Reply #27 on: January 12, 2018, 03:33:03 pm »

Without quoting snippets from earlier posts…

Curiously, it’s previously been an issue with the opposite side of the spectrum with photographic materials being made too greenish-yellow from profiles. My guess is the profile was then attempting to compensate for a rather low yellow gamut (expressed as “Kodak-yellow”) and thus using green to make a more lemon-yellow. Modern inkjet technology doesn’t seem to have as much trouble with yellow nowadays.

Although the apparel industry use spectrophotometers, it's common for the mfr to include an actual fabric swatch to their print provider for color match because they know lighting, surrounds, and sample size matter. 15-20 years ago the apparel industry seemed to more disciplined than they are currently. They would only design with a color their promotional materials could reproduce including their website shopping carts. Seems lately, no so much. I have a friend who operates two UPS stores and he tells me besides fit, color is a common reason for the return by customer to vendor.

If you mouse over a deep blue and observe the CMYK info readout, there is usually a considerable amount of magenta in the mix.

If you have a Kodak color print viewing filter kit and look at the blue card, one might say it appears to be rather purple-ish.

I know many pressmen refer to cyan ink as blue and magenta ink as red. Try printing bills balls without an ICC profile.

I believe nobody hates a blue sky, except for landscape photographers on particular occasions and printers because it reveals how their machines are severely banding and…
Logged

Andy_S

  • Newbie
  • *
  • Offline Offline
  • Posts: 14
Re: Why are dark blues purplish in test print
« Reply #28 on: January 12, 2018, 03:41:49 pm »

Mark, attached are scans of the printed test image with no colour management and with the Ilford GFS profile.  Also a scan of the scanner settings is attached.  Color Control was on.  The scans colours are not the same as the prints.  They are not as bright and the purplish blue is not as intense.  Hope these are helpful.

Wayne, I will print the test image on the Epson Ultra Premium Luster Photo Paper next week when my order arrives.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Why are dark blues purplish in test print
« Reply #29 on: January 12, 2018, 04:09:51 pm »

Thanks Andy. Hard to say much if you think the scans don't adequately portray what the prints look like. But from what I see here: (1) very little difference; (2) the profiled photo is a tad darker all-round than the printer-managed one, so the deeper blue could begin to look a bit purplish but not much; (3) minimal difference for the lapis jewelry; (4) a tad purplish in the row of blue boxes under "Color Match RGB" and in the Blue portion of the rainbow running across the top; (5) neutrals look pretty well preserved in both, except in several of the mid-tone boxes where there appears to be a very slight departure toward greenish/yellow. I'm wondering - if you were to view the target in softproof mode with the IGFS profile active, and in Lr, just lighten the blues a little in the Luminance tab and print it, whether this would largely eliminate the problem in the print. If it does, it could indicate that the profile is rendering blue luminance a bit too dark, causing the appearance of purplish blue.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Why are dark blues purplish in test print
« Reply #30 on: January 12, 2018, 08:00:08 pm »

Andy_S,

One thing that might help is to look at the prints and see which of the following clips of the beads hanging on the model's hand most closely matches. The left most is from the Andrew's image, and each of the following clips has the hue angle increased 5 degrees.

Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Why are dark blues purplish in test print
« Reply #31 on: January 13, 2018, 09:05:19 am »

.................

Badly needed are physical, and accurately reproduced, reference images printed using Rel. Col. that can be purchased or distributed as samples with new printers, combined with the downloadable digital images. ....................

Just reviewing this thread - I forgot to respond to this point you made: Indeed, yes. Such already exists somewhat in the manner you describe: the set of <Roman 16> images from BVDM (Bundesverband Druck und Medien e.V.), Wiesbaden Germany (www.roman16.com). Excellent product, but unfortunately the set costs about 300 EUROS, so not something the printer manufacturers are likely to bundle into their printers at the going printer prices.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

oldshadow

  • Newbie
  • *
  • Offline Offline
  • Posts: 18
Re: Why are dark blues purplish in test print
« Reply #32 on: January 13, 2018, 09:56:10 am »

I'am also seeing the purple blue in this image printed on a P800, Win10 with  the Ilford profile.
Printing from LR, I made a Hue adjustment blue -7 on a virtual copy and the print is now a good visual match to the original.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Why are dark blues purplish in test print
« Reply #33 on: January 13, 2018, 10:16:39 am »

The Roman 16's are an invaluable tool!
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Why are dark blues purplish in test print
« Reply #34 on: January 13, 2018, 10:55:53 am »

Just reviewing this thread - I forgot to respond to this point you made: Indeed, yes. Such already exists somewhat in the manner you describe: the set of <Roman 16> images from BVDM (Bundesverband Druck und Medien e.V.), Wiesbaden Germany (www.roman16.com). Excellent product, but unfortunately the set costs about 300 EUROS, so not something the printer manufacturers are likely to bundle into their printers at the going printer prices.

The Roman 16's are an invaluable tool!

For those that have I1Profiler there is a set of Roman 16 images in the installation. It's in the i1profiler\preview-images in the installation directory.

One can download i1profiler from the xrite website. It won't be enabled w/o an xrite license but the image files are still available.

« Last Edit: January 13, 2018, 11:00:48 am by Doug Gray »
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Why are dark blues purplish in test print
« Reply #35 on: January 13, 2018, 12:01:44 pm »

For those that have I1Profiler there is a set of Roman 16 images in the installation. It's in the i1profiler\preview-images in the installation directory.

One can download i1profiler from the xrite website. It won't be enabled w/o an xrite license but the image files are still available.

Nice one Doug! The images are indeed within the "Package" contents, but the route to finding them is a bit complex. I drilled down using "Pacifist" and am attaching a screen grab of the route to the location of interest to those on OSX. It starts from the unpacking of the installation Package. What one misses in this little gift, however, is the printed material in the product set, with all the technical information about them and showing on paper the correct printed colour appearance for each image - rather important.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Why are dark blues purplish in test print
« Reply #36 on: January 13, 2018, 12:12:33 pm »

For those that have I1Profiler there is a set of Roman 16 images in the installation.
VERYLow rez, (aprox 800x500 pixels) 8-bit per color in wide gamut space; not that useful really.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Why are dark blues purplish in test print
« Reply #37 on: January 13, 2018, 12:21:03 pm »

What one misses in this little gift, however, is the printed material in the product set, with all the technical information about them and showing on paper the correct printed colour appearance for each image - rather important.

Indeed.  Most of the images skirt the boundaries of printer capabilities but without going too far so most printers, properly profiled, should render quite similar images. It's particularly good because the smooth gradients near these boundaries provide an excellent, visual, profile check. It's reduced rez but is still quite good for checking profile performance over the smooth gradients.

Significant amount of money to get the printed material and CD but I believe this is the gold standard for people in the color critical pre-press biz.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Why are dark blues purplish in test print
« Reply #38 on: January 13, 2018, 01:31:57 pm »

VERYLow rez, (aprox 800x500 pixels) 8-bit per color in wide gamut space; not that useful really.

Yup 1.9 MB per photo in i1Profiler versus 137 MB per photo in the original suite.

I would also be concerned about 8-bit colour in a wide gamut space - especially if they are being used for assessing smoothness of tonal gradations.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Why are dark blues purplish in test print
« Reply #39 on: January 13, 2018, 02:34:17 pm »

Yup 1.9 MB per photo in i1Profiler versus 137 MB per photo in the original suite.

I would also be concerned about 8-bit colour in a wide gamut space - especially if they are being used for assessing smoothness of tonal gradations.

It was also highly sharpened when downsized. Very distracting in the fine structure when printed larger than 4x6" or even there looked at with a loupe.  However, the color space, egiRGB v2,  is only slightly larger than Adobe RGB and it is in tif so no jpeg artifacts which is the main gradient banding culprit.  8 bits are just fine and results in no visible impact on smoothness in the larger gradients where defects in profiles show up. Looks nice with them all uprezed to printer native and printed side by side on a larger sheet and it's easy to check the more extreme colors. Tone curve is in L* so the max error for the B&Ws is .2 dE76. Somewhat larger for the color images. Similar to Adobe RGB 8 bit tifs, about .4 dE76 max.
Logged
Pages: 1 [2] 3   Go Up