Pages: 1 2 [3] 4 5   Go Down

Author Topic: Print color frustration  (Read 4688 times)

mdijb

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 666
    • mdiimaging.com
Re: Print color frustration
« Reply #40 on: January 21, 2019, 01:02:23 pm »

Thanks.

I have printed on epson premium luster paper and got the same ugly results.

MDIJB
Logged
mdiimaging.com

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Print color frustration
« Reply #41 on: January 21, 2019, 01:09:12 pm »

That's useful information. Awaiting the rest.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

mdijb

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 666
    • mdiimaging.com
Re: Print color frustration
« Reply #42 on: January 21, 2019, 01:16:14 pm »

Image and profile sent.

MDIJB
Logged
mdiimaging.com

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Print color frustration
« Reply #43 on: January 21, 2019, 01:21:46 pm »

Yup - got them. BTW, that dunes photo is really nice. I can see why you would be concerned! I'll revert once I've done some stuff on the photo and the profile.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Print color frustration
« Reply #44 on: January 21, 2019, 02:06:48 pm »

OK, I think this is becoming quite clear. According to what I see from ColorThink Pro (CTP) you have produced an unprintable photograph. Please see the accompanying screen grab of the CTP analysis. The original pixel size of your photo weighs-in at 166 million. For CTP to work with this, I had to reduce it to 8 bit, then resample it down to 10 x 6.6 inches at 240 PPI using BiCubic Sharper, giving me a revised image size of 16M. That worked fine, the two versions look the same (see accompanying illustration) and CTP can handle this much image data. So I could open the reduced one in CTP. All the gorgeous colors you see there is CTP's extraction and reproduction of every pixel in your photo. Mapped against that are two profiles' gamut volumes and shapes: one for breathing color metallic (the gray blob) and the other for Ilford Gold Fibre Silk (the wire frame) which has an even larger gamut than Epson Premium Luster. You can see from this illustration, that most of the colours in your photo are so far out of the gamut of ANY paper profile that the colour management system would have a hard time dealing with this, and therefore hue shifts have occurred. You could try playing between Relative and Perceptual Rendering Intent, but not at all clear it would make that much difference.

You need to perform less radical editing, doing it under softproof, to see as you go in real time what the printer/paper combination can handle; see whether approaching it that way produces both a good photo and correct colour for you.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Print color frustration
« Reply #45 on: January 21, 2019, 02:10:33 pm »

PS. To Andrew Rodney - I was surprised to see the extent of the colour spread here. Should we have confidence that the CTP rendition is OK? Have you seen such cases before where wildly OOG colours produce hue shifts in a print that don't show on the display? The profile seems well-formed just to look at it. Could it be problems of disconnect between the B>A and A>B tables? (I haven't tried printing it with my own very good profiles - perhaps I should try one on IGFS and see what happens.
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: Print color frustration
« Reply #46 on: January 21, 2019, 02:20:09 pm »

PS. To Andrew Rodney - I was surprised to see the extent of the colour spread here. Should we have confidence that the CTP rendition is OK? Have you seen such cases before where wildly OOG colours produce hue shifts in a print that don't show on the display? The profile seems well-formed just to look at it. Could it be problems of disconnect between the B>A and A>B tables? (I haven't tried printing it with my own very good profiles - perhaps I should try one on IGFS and see what happens.
I have compete faith in your gamut plots. A bit on the road today so didn't download anything; what working space was used?
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Print color frustration
« Reply #47 on: January 21, 2019, 02:23:30 pm »

We're OK there - source space is ProPhoto.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Rand47

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1882
Re: Print color frustration
« Reply #48 on: January 21, 2019, 02:42:19 pm »

OK, I think this is becoming quite clear. According to what I see from ColorThink Pro (CTP) you have produced an unprintable photograph. Please see the accompanying screen grab of the CTP analysis. The original pixel size of your photo weighs-in at 166 million. For CTP to work with this, I had to reduce it to 8 bit, then resample it down to 10 x 6.6 inches at 240 PPI using BiCubic Sharper, giving me a revised image size of 16M. That worked fine, the two versions look the same (see accompanying illustration) and CTP can handle this much image data. So I could open the reduced one in CTP. All the gorgeous colors you see there is CTP's extraction and reproduction of every pixel in your photo. Mapped against that are two profiles' gamut volumes and shapes: one for breathing color metallic (the gray blob) and the other for Ilford Gold Fibre Silk (the wire frame) which has an even larger gamut than Epson Premium Luster. You can see from this illustration, that most of the colours in your photo are so far out of the gamut of ANY paper profile that the colour management system would have a hard time dealing with this, and therefore hue shifts have occurred. You could try playing between Relative and Perceptual Rendering Intent, but not at all clear it would make that much difference.

You need to perform less radical editing, doing it under softproof, to see as you go in real time what the printer/paper combination can handle; see whether approaching it that way produces both a good photo and correct colour for you.

Holy Cats!  No wonder it won't print. I don't think I've ever seen anything quite like this "spread" of color data in ColorThink.  Perhaps the OP could use this rendering as a reference image, and then reprocess using only curves, some local luminance adjustments, and global adjustments that don't include 100% on the contrast slider.  In my experience the contrast slider not only increases contrast but can send saturation off into the stratosphere if pushed this hard.

Rand
Logged
Rand Scott Adams

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Print color frustration
« Reply #49 on: January 21, 2019, 02:58:16 pm »

Well, it depends on what printer and paper you are using as well. I decided to make a print of this photo doing no adjustments whatsoever. Using an Epson SC-P5000 printer and Ilford Gold Fibre Silk paper with my custom profile (gamut volume 985K), Relative Rendering Intent with Black Point Compensation, it produced what I would consider a fine print faithful to a soft-proof of the original, as you can judge from the scanned JPG version of it I prepared for posting here. MDIJB's printer/paper combination has a gamut volume of only 673K, a difference of over 300K, which makes a huge difference in how OOG situations would be handled. I am getting no such hue shift in the yellows as MDIJB reported in his opening post.
(edit for typo)
« Last Edit: January 21, 2019, 03:25:28 pm by Mark D Segal »
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Print color frustration
« Reply #50 on: January 21, 2019, 02:59:16 pm »

You need to click on the image to see the colours correctly. The thumbnail is not colour managed in this Forum structure.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Rand47

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1882
Re: Print color frustration
« Reply #51 on: January 21, 2019, 03:22:36 pm »

Well, it depends on what printer and paper you are using as well. I decided to make a print of this photo doing no adjustments whatsoever. Using an Epson SC-P5000 printer and Ilford Gold Fibre Silk paper with my custom profile (gamut volume 985K), Relative Rendering Intent with Black Point Compensation, it produced what I would consider a fine print faithful to a soft-proof of the original, as you can judge from the scanned JPG version of it I prepared for posting here. MDIJB's printer/paper combination has a gamut volume of only 673K 9 difference of over 300K, which makes a huge difference in how OOG situations would be handled. I am getting no such hue shift in the yellows as MDIJB reported in his opening post.

Mark,

That's pretty amazing.  Might have to try it myself on my P5000 w/ a custom profile I have for GFS.  What does it look like when you hover over the gamut warning in the soft proof?  (I know that doesn't tell you "how much out of gamut," but thought it might be "telling" in terms of what one is dealing with.)  :-)

Rand
Logged
Rand Scott Adams

Rand47

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1882
Re: Print color frustration
« Reply #52 on: January 21, 2019, 03:27:21 pm »

You need to click on the image to see the colours correctly. The thumbnail is not colour managed in this Forum structure.

Mark,

Looking at the ColorThink plot, do you think that using perceptual rendering intent might be the source of moving all that "not visible" green into the paper's gamut?  Relative just chopped it off clean?  Trying to learn from this, and sometimes my knowledge is "just enough" to get me into trouble in my conceptual thinking about this stuff.

Rand
Logged
Rand Scott Adams

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Print color frustration
« Reply #53 on: January 21, 2019, 03:31:38 pm »

Rand,

That Gamut Warning really isn't too useful so I tend not to bother with it. But just to answer your question, it flicked up some black here and there. I did play back and forth with Relative and Perceptual Intents before printing and saw under softproof that Relative would provide a more faithful result, but the difference wasn't that dramatic.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

mdijb

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 666
    • mdiimaging.com
Re: Print color frustration
« Reply #54 on: January 21, 2019, 03:50:49 pm »

WOW!  The print you made looks just like what I wanted.  Thanks for diagnosing the problem and all the work you did.

Are the profiles supplied by the paper makers adequate to do this job?  I recently tried a sample of Epson Baryta and liked what  I saw.

MDIJB
Logged
mdiimaging.com

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Print color frustration
« Reply #55 on: January 21, 2019, 04:05:20 pm »

You are welcome.

Manufacturers' profiles vary, but definitely an improving story over the years. For example, Epson's for the Legacy Baryta paper are good.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Alan Goldhammer

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4344
    • A Goldhammer Photography
Re: Print color frustration
« Reply #56 on: January 21, 2019, 04:50:05 pm »

Good job solving the OP's problem.  This is why LuLa is a valuable resource for us all!!!
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Print color frustration
« Reply #57 on: January 21, 2019, 05:06:54 pm »

Are the profiles supplied by the paper makers adequate to do this job?  I recently tried a sample of Epson Baryta and liked what  I saw.



Not all ICC profiles are created equally

In this 23 minute video, I'll cover:
The basic anatomy of ICC Profiles
Why there are differences in profile quality and color rendering
How to evaluate an ICC output profile
Examples of good and not so good canned profiles and custom profiles on actual printed output.

High resolution: http://digitaldog.net/files/Not_All_Profiles_are_created_equally.mp4
Low resolution (YouTube): https://www.youtube.com/watch?v=TNdR_tIFMME&feature=youtu.be
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Stephen Ray

  • Full Member
  • ***
  • Offline Offline
  • Posts: 217
Re: Print color frustration
« Reply #58 on: January 21, 2019, 05:07:32 pm »

OK, I think this is becoming quite clear. According to what I see from ColorThink Pro (CTP) you have produced an unprintable photograph. Please see the accompanying screen grab of the CTP analysis. The original pixel size of your photo weighs-in at 166 million. For CTP to work with this, I had to reduce it to 8 bit, then resample it down to 10 x 6.6 inches at 240 PPI using BiCubic Sharper, giving me a revised image size of 16M. That worked fine, the two versions look the same (see accompanying illustration) and CTP can handle this much image data. So I could open the reduced one in CTP. All the gorgeous colors you see there is CTP's extraction and reproduction of every pixel in your photo. Mapped against that are two profiles' gamut volumes and shapes: one for breathing color metallic (the gray blob) and the other for Ilford Gold Fibre Silk (the wire frame) which has an even larger gamut than Epson Premium Luster. You can see from this illustration, that most of the colours in your photo are so far out of the gamut of ANY paper profile that the colour management system would have a hard time dealing with this, and therefore hue shifts have occurred. You could try playing between Relative and Perceptual Rendering Intent, but not at all clear it would make that much difference.

You need to perform less radical editing, doing it under softproof, to see as you go in real time what the printer/paper combination can handle; see whether approaching it that way produces both a good photo and correct colour for you.

A hue shift should not have happened with a good profile. Saturation shift, yes, but not a hue shift.
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Print color frustration
« Reply #59 on: January 21, 2019, 05:10:38 pm »

This thread is an excellent example of the dangers lurking using ProPhoto RGB and adjusting colors to what looks good on a monitor.

OTOH, only ProPhoto RGB will allow you to print all the colors most printer/paper combos can do.

So, with great power comes great responsibility, Apologies to Stan Lee.

ProPhoto RGB has a huge gamut. So large some colors are imaginary. There can be big differences in both how an RGB color is displayed, and how it's printed.

The transforms from RGB spaces goes through a completely different process to the monitor v to the printer. OOG colors will usually be brightened to the monitor because the conversion produces negative numbers on one or more channels. These are clipped to 0 which raises the overall brightness as well as sometimes shifting the hue in addition to the expected clipping of saturation.

On the printer side things are more complicated. RGB colors are converted to a limited LAB space where a* and b* are first clipped to <= 128 magnitude.  Then the profile maps these to somewhere on the printer gamut surface. This process is done w/o any knowledge of how the same RGB values are altered for display.

The general solution is soft proofing because this uses the printer profile to generate, then reverse, the printable colors then converting them to the display RGB colorspace. It's not perfect since some printable colors aren't displayable. Even on wide gamut monitors. These are clipped. But it's a big improvement over not soft proofing.
Logged
Pages: 1 2 [3] 4 5   Go Up