Pages: 1 2 [3]   Go Down

Author Topic: What exactly does 'Simulate paper white' do?  (Read 17544 times)

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20944
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: What exactly does 'Simulate paper white' do?
« Reply #40 on: October 04, 2017, 06:11:54 pm »

Andrew wasn't talking about two profiles, but the two tables in a profile, one of which is used for rendering color to the printer, the other takes the RGB values from that rendering to generate the actual colors printed.
I was talking about one table that is for printing, one for soft proofing. Two tables, one profile.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2205
Re: What exactly does 'Simulate paper white' do?
« Reply #41 on: October 04, 2017, 06:50:49 pm »

I was talking about one table that is for printing, one for soft proofing. Two tables, one profile.

At the risk of going further into geeky detail, in a profile there are normally 3 tables for printing, one each for Perceptual, Relative, and Saturation Intents.  These are called BtoA tables. The inverse Relative table, a AtoB table, is used for creating soft proofs or hard proofs when cross rendering (simulating a print on a different device). It's purpose is to report back, within the limits of interpolation and measurement error, the color printed from whichever one of the printing intents were selected. The Relative Intent table is shared with Absolute Intent in both directions though V4 profiles allow for a separate table. I've not run across a profile that has one.

http://www.color.org/specification/ICC1v43_2010-12.pdf

Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20944
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: What exactly does 'Simulate paper white' do?
« Reply #42 on: October 04, 2017, 06:55:19 pm »

At the risk of going further into geeky detail, in a profile there are normally 3 tables for printing, one each for Perceptual, Relative, and Saturation Intents. 
Indeed yes, but those are lumped into the 'output' table I referring to. And for soft proofing, there would be the same set of tables depending on the profile structure of course.
Your text about my text seemed to aim only towards output at least as I read it:


Quote
but the two tables in a profile, one of which is used for rendering color to the printer, the other takes the RGB values from that rendering to generate the actual colors printed.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2205
Re: What exactly does 'Simulate paper white' do?
« Reply #43 on: October 04, 2017, 07:58:34 pm »

Indeed yes, but those are lumped into the 'output' table I referring to. And for soft proofing, there would be the same set of tables depending on the profile structure of course.
Your text about my text seemed to aim only towards output at least as I read it:

Good point. I skipped the step where the rgb values sent to the device driver are converted back to Lab (the PCS), an intermediate point which is then processed by the monitor's profile to produce the proof display. It is worth it to take time to go through what happens behind the scenes and it does help for folks to understand just how difficult it is to soft proof when OBAs are present. Heck, it isn't even that easy without OBAs complicating things but it is doable and it is worth it.

Andrew, your videos are really excellent vehicles for people just wading into color management. They include the necessary background and detail without anything but the necessary tech speak. The resources at color.org are also excellent but can be intimidating for many. They are great for those that want to delve into the gory details later.
Logged

Peter_DL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 544
Re: What exactly does 'Simulate paper white' do?
« Reply #44 on: October 07, 2017, 06:20:43 am »

I still don't understand in which way the soft proof should match the monitor profile.

One philosophy is that the monitor gets calibrated to a Brightness and a White point that visually matches the final output i.e. the print.

There were some earlier discussions and different opinions, if this matching and the comparison screen/print should be done under Soft-proof with or without Simulate paper color.

In any case it requires to be consistent then, when soft-proofing in practice.

Gruß, Peter

Simulate Paper Color:
https://creativepro.com/out-of-gamut-photoshop-previsualization-and-print-prediction/
http://www.cambridgeincolour.com/tutorials/soft-proofing.htm
Logged

Hening Bettermann

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 945
    • landshape.net
Re: What exactly does 'Simulate paper white' do?
« Reply #45 on: October 07, 2017, 01:50:05 pm »

Thank you Peter for these 2 very useful links!

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20944
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: What exactly does 'Simulate paper white' do?
« Reply #46 on: October 07, 2017, 07:54:38 pm »

One philosophy is that the monitor gets calibrated to a Brightness and a White point that visually matches the final output i.e. the print.
There were some earlier discussions and different opinions, if this matching and the comparison screen/print should be done under Soft-proof with or without Simulate paper color.


Exactly!
With simulation.
One experts opinion about cambridgeincolor.com:
https://www.dpreview.com/forums/post/60217235
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2205
Re: What exactly does 'Simulate paper white' do?
« Reply #47 on: October 07, 2017, 10:10:53 pm »


Exactly!
With simulation.
One experts opinion about cambridgeincolor.com:
https://www.dpreview.com/forums/post/60217235

I wonder how many people reading that link realize the image they incorrectly tout as the "Profile Connection Space" is actually the human gamut CIEuv chromaticity graph?  Might as well have a cat picture.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20944
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: What exactly does 'Simulate paper white' do?
« Reply #48 on: October 07, 2017, 11:13:15 pm »

I wonder how many people reading that link realize the image they incorrectly tout as the "Profile Connection Space" is actually the human gamut CIEuv chromaticity graph?  Might as well have a cat picture.
ROTFL!
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Peter_DL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 544
Re: What exactly does 'Simulate paper white' do?
« Reply #49 on: October 08, 2017, 06:39:30 am »

One philosophy is that the monitor gets calibrated to a Brightness and a White point that visually matches the final output i.e. the print.
There were some earlier discussions and different opinions, if this matching and the comparison screen/print should be done under Soft-proof with or without Simulate paper color.
In any case it requires to be consistent then, when soft-proofing in practice.

Exactly!
With simulation.

with or without Simulate paper color:
http://forum.luminous-landscape.com/index.php?topic=46010.msg385595#msg385595
http://forum.luminous-landscape.com/index.php?topic=47185.0

The general problem lies in the RelCol-mapping of the source white onto the target white. To assume a 100% Chromatic adaptation to the target white (RelCol WP mapping) might be as wrong as to assume zero Chromatic adaptation to the target white (AbsCol WP mapping). In reality, there is adaptation, but it is not complete (see pdf, chapter 5).

The solution to visually adjust the monitors calibrated white seems to me somewhat ironic/paradox, considering all the science & geek stuff which comes along with Color management.

Peter

--
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2205
Re: What exactly does 'Simulate paper white' do?
« Reply #50 on: October 08, 2017, 01:15:54 pm »

with or without Simulate paper color:
http://forum.luminous-landscape.com/index.php?topic=46010.msg385595#msg385595
http://forum.luminous-landscape.com/index.php?topic=47185.0

The general problem lies in the RelCol-mapping of the source white onto the target white. To assume a 100% Chromatic adaptation to the target white (RelCol WP mapping) might be as wrong as to assume zero Chromatic adaptation to the target white (AbsCol WP mapping). In reality, there is adaptation, but it is not complete (see pdf, chapter 5).

The solution to visually adjust the monitors calibrated white seems to me somewhat ironic/paradox, considering all the science & geek stuff which comes along with Color management.

Peter

--

Adaptation is an incredibly complex topic and one I avoid. I run my monitor at D50 and I use ProPhoto which has RGB coordinates at D50 unlike most, but not all, other standard RGB spaces. There are other issues as well including surround variation between the monitor and print viewing station. There are ways to mitigate and minimize those differences but that's only part of the problem.

The larger problem I have is the difference between 2 degree and 10 degree CIEXYZ with my two Eizo monitors. Both are wide gamut but one has a CFL backlight, the other an LED backlight. The peaks of the blues for the two monitors differ by almost 20nm.  And, as luck would have it, that has a big impact on the shift in white points between wide area (10 degree)  and spot colors (2 degree). Thus, if I profile both for the same white point CIExy (2 degrees) which is the standard, the screen whites, which are wide area, have different tints. The LED Eizo has the least shift based on its spectrum and the CIE 10/2 responses and that's the one I use for proofing. I adjusted the white point on the CFL Eizo away from D50 such that the wide angle white matches the LED Eizo. I don't use that one for color work but at least I can set its white point so that it matches visually and isn't a distraction.

Another issue is the variation human to human in color response. This is an issue with monitors generally compared to prints. Some research indicates humans can be upwards of 10 dE from one person to another and there really is no way for any one person to evaluate how close their color perception is to the CIE's Standard Observer.
Logged

Hening Bettermann

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 945
    • landshape.net
Re: What exactly does 'Simulate paper white' do?
« Reply #51 on: November 05, 2017, 10:12:50 am »

Round up:

In this thread as well as in an earlier one
http://forum.luminous-landscape.com/index.php?topic=120611.0
I asked because I had problems with a print. The highlights seemed dull, and I learned that I had to apply more light for viewing. The print also seemed over all yellowish, and I suspected problems with my soft proofing and OBAs, or thought the failure was on the side of the print service. All wrong! Ernst Dinkla (pigment-print.com) made some proofs for me, and it turned out that both his and those from my first print service (digitaloriginal.de) were perfect in both contrast and color - if viewed under a sufficient amount of light!

And my question why the soft proof showed a difference in contrast, but not in color, may have a very simple answer: The colors of my image were all inside the paper gamut! (The image contains no deep blues or purples).

Thanks again to all who chimed in!

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2205
Re: What exactly does 'Simulate paper white' do?
« Reply #52 on: November 05, 2017, 03:42:27 pm »

Andrew
>Paper White! Now that isn't to say the profile is 'correct' (we'd have to do some measurements) or your display white point isn't nailed but the profile predicts a warm paper white.
TonyW
>This may indeed be the case for the individual user and if I was to think of a reason or two why this should happen my thoughts would first stray to the validity of monitor calibration and then the quality/accuracy of the ICC profile for a particular paper and viewing/lighting conditions for the print. 

Hm. How could I know the validity of my monitor calibration? There is an option to validate the profile, using the same software that created it, so how is the validation of this validation? Andrew has earlier called it a 'Feel-good-button'. How can I check the white point of the monitor profile?
Viewing conditions for the print: I have calibrated the monitor to D50 and have established D50 roomlight not Solux but Yuji,
 
https://store.yujiintl.com/collections/high-cri-led-lights/products/bc-series-a60-high-cri-remote-phosphor-led-bulb-unit-2-pcs?variant=27740384775

even though not enough of it.

The print looks rather yellow/brown, like it had spent to much time in the fixation bath. I'll try to simulate the amount on screen when I'm back home in some days, I'm on the road right now.
The paper in question is Epson Premium Glossy. It does contain OBAs, not sure how much of them.

I have some Epson Prem. Glossy. It doesn't have OBAs on the print surface but the substrate is fairly fluorescent. The paper I have is just barely bluish w/o uV, Lab=95,-1,-1  (M2) and only slightly more, Lab=95,0,-3, with D50 (M1).

Not sure what you mean by prints being yellowish. Does the unprinted paper appear yellowish? One way to compare to a good white is to get a roll of plumber's PTFE tape. It's very cheap stuff and really thin designed to seal pipe threads but has a matte surface and but reflects light evenly and is spectrally flat. Lay a piece of  across unprinted paper to see what kind of tint a paper has.

You can rough check the paper's profile for this sort of "tint" by printing a colorchecker (download from babelcolor) using Absolute Colorimetric intent. The print should match a colorchecker quite well when viewed in daylight. If not, you have a profile/printer/paper or settings mismatch.

As for "show paper white" the paper is fairly neutral with only a small amount of fluorescence but with a very slight blue shift. So you should also see a very slight blue shift with "show paper white."  However, you will most likely notice the much larger luminance decrease because the L*=95 is a significant drop from pure white.
Logged

Hening Bettermann

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 945
    • landshape.net
Re: What exactly does 'Simulate paper white' do?
« Reply #53 on: November 05, 2017, 05:08:34 pm »

Hello Doug Gray

> Not sure what you mean by prints being yellowish. Does the unprinted paper appear yellowish?

No, it was not the unprinted paper white (which was not visible anywhere in the image). The print as a whole looked yellowish/brown, in particular visible in the greens (of which the image has a lot). Ernst Dinkla has written to me that "The Kruithof curve learns us that humans have a preference for warmer light when the light level drops. All to do with cons and rods in the eye." So this is the reason - in a sufficient amount of light the print looks perfect. Nothing wrong with print service, paper, profiles or soft proof.

Thanks again!
Pages: 1 2 [3]   Go Up