Pages: 1 2 [3] 4 5   Go Down

Author Topic: No color management  (Read 5084 times)

texshooter

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 458
Re: No color management
« Reply #40 on: February 13, 2018, 09:13:30 PM »

So, what happens when you make prints with PROFILES ON? Do you see the results you're looking for?

When I print the ProphotoRGB target using the ICC profile I made by the ColorMunki, I get *L 57.6  (should be *L 61 according to the info panel in PS)
When I print the AdobeRGB target using the ICC profile (same profile as above), I get *L 51.3  (Should be *L 54 according to the info panel in PS)

This tells me that although the ICC profile has missed its mark, it at least can tell the difference between a ProphotoRGB image (gamma 1.8] and a AdobeRGB image (gamma 2.2). I deduce this because the measured Lab value of the ProphotoRGB print is roughly 6 points higher than the AdobeRGB print, which is in the right direction, thank God.

The how is a mystery.

Sidebar:

Because I don't use an ICC profile when printing to the Epson ABW driver, I was curious to see how the ABW driver would render these two targets.

The ProphotoRGB target printed with *L 52.0  And the AdobeRGB target printed with *L 52.1  Exactly the same!  This tells me that the ABW driver converted the ProPhotoRGB image into AdobeRGB. Or more precisely, gamma 1.8 to gamma 2.2.    That's why they say always to work in gamma 2.2 when planning to print to the ABW driver. You'll never be able to match your print to screen otherwise.

« Last Edit: February 13, 2018, 09:18:19 PM by texshooter »
Logged

Doug Gray

  • Sr. Member
  • ****
  • Online Online
  • Posts: 1237
Re: No color management
« Reply #41 on: February 13, 2018, 09:14:29 PM »

PS:  Yes, your printer is way out of calibration.

No, it's not. Those Lab numbers texshooter measured are not unreasonable. Apart from that printer calibration is something one does with a RIP. RGB calibration requires a service manual and the purpose is to bring the printer into spec so that the canned profiles provided produce accurate prints. It's not something accessible to end users. It's really for things like having the head changed by a manufacturing trained service tech and even then many don't do it if they even know how. If you make your own profiles it really isn't necessary unless you use a RIP which directly controls the CYMK inking. Calibration with a RIP allows you to skip having to create new profiles when the printer drifts.
Logged

Doug Gray

  • Sr. Member
  • ****
  • Online Online
  • Posts: 1237
Re: No color management
« Reply #42 on: February 13, 2018, 09:21:30 PM »

When I print the ProphotoRGB target using the ICC profile I made using the ColorMunki, I get *L 57.6  (should be *L 61 according to the info panel in PS)
When I print the AdobeRGB target using the ICC profile (same profile as above), I get *L 51.3  (Should be *L 54 according to the info panel in PS)

Au contraire. That's about what you should get. Printed colors are scaled to the paper's white point. If you measure the unprinted paper L* it will be under 100, typically about 95 or 96.  If L* wasn't scaled then high key prints would have their brightest areas clipped to the paper's white point. Not good. If you want to print a specific color exactly, and it's in the printer's gamut, print using Absolute Colorimetric. You will find the L* values increases and will be quite close to the L* you get reading the info in Photoshop.
Logged

texshooter

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 458
Re: No color management
« Reply #43 on: February 13, 2018, 09:32:52 PM »

Au contraire. That's about what you should get. Printed colors are scaled to the paper's white point. If you measure the unprinted paper L* it will be under 100, typically about 95 or 96.  If L* wasn't scaled then high key prints would have their brightest areas clipped to the paper's white point. Not good. If you want to print a specific color exactly, and it's in the printer's gamut, print using Absolute Colorimetric. You will find the L* values increases and will be quite close to the L* you get reading the info in Photoshop.

Ah, je vois.
Logged

GWGill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 452
  • Author of ArgyllCMS & ArgyllPRO ColorMeter
    • ArgyllCMS
Re: No color management
« Reply #44 on: February 13, 2018, 10:11:13 PM »

My ColorMunki only came with one button, not two, so I'd be interested to know what device you use to calibrate with and what device you use to characterize with.
OK - so you are a troll. No point in responding any further then.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 11278
    • http://www.markdsegal.com
Re: No color management
« Reply #45 on: February 13, 2018, 10:21:18 PM »

OK - so you are a troll. No point in responding any further then.

According to Wiki: In Internet slang, a troll (/troʊl, trɒl/) is a person who sows discord on the Internet by starting quarrels or upsetting people, by posting inflammatory,[1] extraneous, or off-topic messages in an online community (such as a newsgroup, forum, chat room, or blog) with the intent of provoking readers into an emotional response[2] or of otherwise disrupting normal, on-topic discussion,[3] often for the troll's amusement.

Now, this poster is a Senior Member who joined in 2011 with no particular history of "trolling", so not clear to me why you are saying this.

I think this person is trying to understand certain principles of how colour management works and he's not there yet despite some remarkably good responses, including your technical ones, in this thread.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

nirpat89

  • Full Member
  • ***
  • Offline Offline
  • Posts: 213
    • Photography by Niranjan Patel
Re: No color management
« Reply #46 on: February 13, 2018, 10:41:43 PM »

Wow this things still going on...even after the excellent description of kirkt.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 11278
    • http://www.markdsegal.com
Re: No color management
« Reply #47 on: February 13, 2018, 10:56:16 PM »

Reply 35 seems to indicate a basic misunderstanding of what matters for the production of targets intended to be used for measurement and the creation of paper/printer profiles. The device independent L*a*b* values in the target reference file are the operational values. These target file values should be printed with no embedded RGB profile. The printer reproduces them in "printer space" thereby characterizing how the printer reproduces those target file values. The spectro is used to read the printed target and create the profile as kirkt described it in his very post above. There is no need and no point to messing with RGB values in targets through this whole procedure.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Lundberg02

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 376
Re: No color management
« Reply #48 on: February 14, 2018, 12:02:20 AM »

three pages of restating the obvious, when the OP could have googled it
Logged

texshooter

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 458
Re: No color management
« Reply #49 on: February 14, 2018, 12:12:02 AM »

Reply 35 seems to indicate a basic misunderstanding of what matters for the production of targets intended to be used for measurement and the creation of paper/printer profiles.

I'm going to assume that all printers are programmed from the factory to render RGB 128,128,128 as a shade of gray that is exactly 50% as dark as the darkest shade the printer is capable of printing.  If zero picoliters of ink quantifies the lightest shade the printer nozzle can make, and if 100 picoliters of ink quantifies the darkest shade the printer nozzle can make, then RGB 128,128,128 will land on the page as 50 picoliters of ink.  Without color management, of course.

Whether true or not, this visual helps me.
Logged

texshooter

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 458
Re: No color management
« Reply #50 on: February 14, 2018, 12:14:03 AM »

three pages of restating the obvious, when the OP could have googled it

You presume to understand my question as I intended it.
Logged

Doug Gray

  • Sr. Member
  • ****
  • Online Online
  • Posts: 1237
Re: No color management
« Reply #51 on: February 14, 2018, 12:40:09 AM »

I'm going to assume that all printers are programmed from the factory to render RGB 128,128,128 as a shade of gray that is exactly 50% as dark as the darkest shade the printer is capable of printing.

There is no particular reason for factories to do this. There is only need for consistency for any given printer model. Perhaps in the earliest days of printers before color management was codified something like that occurred.
Logged

texshooter

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 458
Re: No color management
« Reply #52 on: February 14, 2018, 12:49:10 AM »

There is no particular reason for factories to do this. There is only need for consistency for any given printer model. Perhaps in the earliest days of printers before color management was codified something like that occurred.

Since RGB 128,128,128 always prints the same (with CM off) by the same printer. I cant see how the printer knows how dark to render 128,128,128 unless the manufacturer programmed the printer to spit out a predetermined volume of ink for each 128,128,128 pixel.  No?
Logged

Doug Gray

  • Sr. Member
  • ****
  • Online Online
  • Posts: 1237
Re: No color management
« Reply #53 on: February 14, 2018, 12:59:30 AM »

Since RGB 128,128,128 always prints the same (with CM off) by the same printer. I cant see how the printer knows how dark to render 128,128,128 unless the manufacturer programmed the printer to spit out a predetermined volume of ink for each 128,128,128 pixel.  No?
Of course. A manufacturer designs a printer to produce a certain level of repeatability for any given RGB triplet. They just don't much care about forcing it to a specific value when color management is disabled.  It can and does vary with printers and even different paper settings on the same printer.
Logged

nirpat89

  • Full Member
  • ***
  • Offline Offline
  • Posts: 213
    • Photography by Niranjan Patel
Re: No color management
« Reply #54 on: February 14, 2018, 01:54:04 AM »

Of course. A manufacturer designs a printer to produce a certain level of repeatability for any given RGB triplet. They just don't much care about forcing it to a specific value when color management is disabled.  It can and does vary with printers and even different paper settings on the same printer.
The printer manufacturer provides accuracy.  Color management provides precision

Yes, no, may be?

_________

Correction:  Should be the other way around.  Caught by kirkt (see post #60 for elaborate explanation.)
« Last Edit: February 14, 2018, 09:19:54 AM by nirpat89 »
Logged

texshooter

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 458
Re: No color management
« Reply #55 on: February 14, 2018, 01:55:57 AM »

Of course. A manufacturer designs a printer to produce a certain level of repeatability for any given RGB triplet. They just don't much care about forcing it to a specific value when color management is disabled.  It can and does vary with printers and even different paper settings on the same printer.

So let's run with that.  Let's say that my specific printer/paper settings combo was preprogrammed to lay down 50 units of ink for every 128,128,128 pixel (with CM off).

Now we feed the printer a sheet of EEF and let Colormunki do its thing. (Actually, two sheets because ColoMunki makes two passes.)

After that is said and done, we now have an ICC profile.

Now let's print out my ProphotoRGB 11-step target with CM turned ON and the above ICC profile selected.  For simplicity sake, let's just print out the zone V patch, the one with an eyedropper value  of RGB 128,128,128 and LAB 61.

If we could examine this print under a microscope, we would no doubt see that the printer did not lay down 50 units of ink like it did when CM was turned off.  On the contrary, the actual number would be greater or fewer than 50 units due to the influence of the ICC profile. Let's say, for the sake of argument, that the ICC profile caused the printer to lay down 40 units of ink instead of 50. Why 40 units, you ask?  Because that's how many units of black ink it took to satisfy the Colormunki. The Colormunki wants middle gray to render a certain way and, by God, it will see it through.

What I want to know is what does the Colormunki think 128,128,128 should look like?  It must have its own definition of middle gray preprogrammed into it by the manufacturer. And this definition of middle gray  must be in units of L*a*b because that's how the Colormunki spectrophotometer see things--in L*a*b.

 It is the nature of this internal preprogrammed definition of middle gray as it exists in the eye of the Colormunki that lies at the heart of my inquiry.

Or is the rabbit hole getting too deep?
« Last Edit: February 14, 2018, 02:16:03 AM by texshooter »
Logged

Doug Gray

  • Sr. Member
  • ****
  • Online Online
  • Posts: 1237
Re: No color management
« Reply #56 on: February 14, 2018, 02:38:15 AM »

So let's run with that.  Let's say that my specific printer/paper settings combo was preprogrammed to lay down 50 units of ink for every 128,128,128 pixel (with CM off).

Now we feed the printer a sheet of EEF and let Colormunki do its thing. (Actually, two sheets because ColoMunki makes two passes.)

After that is said and done, we now have an ICC profile.

Now let's print out my ProphotoRGB 11-step target with CM turned ON and the above ICC profile selected.  For simplicity sake, let's just print out the zone V patch, the one with an eyedropper value  of RGB 128,128,128 and LAB 61.

If we could examine this print under a microscope, we would no doubt see that the printer did not lay down 50 units of ink like it did when CM was turned off.  On the contrary, the actual number would be greater or fewer than 50 units due to the influence of the ICC profile. Let's say, for the sake of argument, that the ICC profile caused the printer to lay down 40 units of ink instead of 50. Why 40 units, you ask?  Because that's how many units of black ink it took to satisfy the Colormunki. The Colormunki wants middle gray to render a certain way and, by God, it will see it through.

What I want to know is what does the Colormunki think 128,128,128 should look like?  It must have its own definition of middle gray preprogrammed into it by the manufacturer. And this definition of middle gray  must be in units of L*a*b because that's how the Colormunki spectrophotometer see things--in L*a*b.

 It is the nature of this internal preprogrammed definition of middle gray as it exists in the eye of the Colormunki that lies at the heart of my inquiry.

Or is the rabbit hole getting too deep?

The Colormunki is a spectrophotometer. It measures the percentage of light that is reflected at regular intervals across the visible spectrum. This data is recorded and mathematically transformed using the CIE 1931 human, 2 degree standard observer model which produces 3 values, called X, Y, and Z. These are then converted to Lab.

You can find all the math details and learn much about colorspaces here:
http://www.brucelindbloom.com/

Enjoying the rabbithole?
Logged

texshooter

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 458
Re: No color management
« Reply #57 on: February 14, 2018, 02:57:10 AM »

You can find all the math details and learn much about colorspaces here:
http://www.brucelindbloom.com/

Enjoying the rabbithole?

Too deep for me.
Logged

Alan Goldhammer

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2688
    • A Goldhammer Photography
Re: No color management
« Reply #58 on: February 14, 2018, 08:18:58 AM »

My ColorMunki only came with one button, not two, so I'd be interested to know what device you use to calibrate with and what device you use to characterize with.
You can use the ColorMunki to both calibrate (if desired) and profile your printer using the ArgyllCMS software that Graeme Gill developed and maintains.  A useful primer, in addition to the materials on the Argyll website can be found at:  https://www.ludd.ltu.se/~torger/photography/argyll-print.html#toc0

Alan
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 11278
    • http://www.markdsegal.com
Re: No color management
« Reply #59 on: February 14, 2018, 08:23:54 AM »

I'm going to assume that all printers are programmed from the factory to render RGB 128,128,128 as a shade of gray that is exactly 50% as dark as the darkest shade the printer is capable of printing.  If zero picoliters of ink quantifies the lightest shade the printer nozzle can make, and if 100 picoliters of ink quantifies the darkest shade the printer nozzle can make, then RGB 128,128,128 will land on the page as 50 picoliters of ink.  Without color management, of course.

Whether true or not, this visual helps me.

I think I mentioned in a post further up that the printers are linearized and calibrated to a common company standard before leaving the factory and we are not made privy to what those standards are, but if you profile the printer and the profiling works properly we don't need to concern ourselves about this. No matter what they do, printer performance can vary from user to user for a number of different reasons and that is why we do custom profiling.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."
Pages: 1 2 [3] 4 5   Go Up