Pages: 1 2 3 [4] 5   Go Down

Author Topic: No color management  (Read 16648 times)

kirkt

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 604
Re: No color management
« Reply #60 on: February 14, 2018, 09:02:08 am »

The printer manufacturer provides accuracy.  Color management provides precision

Yes, no, may be?

The other way around.  The printer manufacturer makes sure that if you feed it an RGB triplet it will print that triplet in a repeatable, well characterized manner so that if you feed it the triplet 100 times, the variation between prints will be small, within some quantifiable tolerance for coverage uniformity, measured reflectance or density, etc.  That is precision, regardless of whether or not that printed value represents what the user desires (the "actual" value).

Accuracy is the printer's ability to print an input that represents a reference value - if you feed the printer an RGB triplet that represents L*50, it should print at L*50.  If it does, it is accurate, if it does not it is not accurate.  Color management and ICC profiles provide the method by which the printer's native color (which is precise and repeatable) gets transformed into an accurate representation of a specific target color - that is, the difference between the print and the reference (usually expressed as deltaE) is within some tolerance.  If the difference between the print and the reference falls within the tolerance, the printer is "accurate."

Think of it this way:  You are sighting in a rifle - you take 5 shots at a target and they are all very tightly clustered (precise) but high and left of the target's bull's eye.  So the rifle is precise (tight cluster of hits) but not accurate (the bull's eye is the reference).  So you adjust the scope via an ICC profile and, when done correctly, the tight cluster of shots is still tightly clustered (precise) but now it is tightly clustered on the bull's eye (accurate).

In this analogy, you do not adjust the rifle (the printer) you adjust the scope (the ICC profile) to optimize the inherent precision of the rifle so that it is both precise AND accurate.

kirk
Logged

nirpat89

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 661
    • Photography by Niranjan Patel
Re: No color management
« Reply #61 on: February 14, 2018, 09:09:04 am »

The other way around.  The printer manufacturer makes sure that if you feed it an RGB triplet it will print that triplet in a repeatable, well characterized manner so that if you feed it the triplet 100 times, the variation between prints will be small, within some quantifiable tolerance for coverage uniformity, measured reflectance or density, etc.  That is precision, regardless of whether or not that printed value represents what the user desires (the "actual" value).

Accuracy is the printer's ability to print an input that represents a reference value - if you feed the printer an RGB triplet that represents L*50, it should print at L*50.  If it does, it is accurate, if it does not it is not accurate.  Color management and ICC profiles provide the method by which the printer's native color (which is precise and repeatable) gets transformed into an accurate representation of a specific target color - that is, the difference between the print and the reference (usually expressed as deltaE) is within some tolerance.  If the difference between the print and the reference falls within the tolerance, the printer is "accurate."

Think of it this way:  You are sighting in a rifle - you take 5 shots at a target and they are all very tightly clustered (precise) but high and left of the target's bull's eye.  So the rifle is precise (tight cluster of hits) but not accurate (the bull's eye is the reference).  So you adjust the scope via an ICC profile and, when done correctly, the tight cluster of shots is still tightly clustered (precise) but now it is tightly clustered on the bull's eye (accurate).

In this analogy, you do not adjust the rifle (the printer) you adjust the scope (the ICC profile) to optimize the inherent precision of the rifle so that it is both precise AND accurate.

kirk

Yes, Kirk:  I thought the right way, wrote the wrong way....thanks for catching the obvious booboo.  Your elaboration is perfect!
Logged

kirkt

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 604
Re: No color management
« Reply #62 on: February 14, 2018, 09:33:43 am »


...

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 has no opinion about how 128, 128, 128 should look.  It is a measuring device.  The output of the system you use to print a target with no color management (the representation of the printer's native output) is measured by the Colormunki and compared to the known target values to form the basis for the ICC profile.  The profiling software is the thing that will compare the target values for each patch (stored in a reference file, usually with a ".cie" file extension) to the measured values you acquired with the Colormunki.

Maybe it is not apparent to you - RGB color spaces are artificial representations of color - you need to know their primaries (chromaticities) and white point to understand how an RGB triplet represents a specific color that we can see.  Because each RGB color space has its own primaries and white point, the color "Cherry Red" might be 235, 15, 35 in one color space and 220, 13, 40 in another - they both are the exact same color, but the RGB triplets are different because they originate in different RGB color models.  This is what makes them DEVICE-DEPENDENT.  In contrast, Profile Connection Spaces (PCS) are based in the physiology of how a "Standard" human observer perceives light of a specific "tristimulus" value - that is, how our visual systems interpret that light.  These physically-based models of color form the basis for our ability to relate various artificial, device-dependent RGB models to each other - as such, they are DEVICE-INDEPENDENT.  Hence the term "profile connection spaces" - the physically-referred PCS can connect one artificial RGB space to another, as long as the relationship between the RGB space and the PCS is known.  It is sort of like Google Translate for color.  Again, I would defer to the color experts here, as I am sure I am oversimplifying.

The printer internals are designed to lay down ink.  How much gets laid down in response to a native RGB input is beyond me, but I assume that printer manufacturers have it figured out pretty well for their ink formulations and media types.  It sounds like you are trying to adopt the Zone System for digital photography and you want to understand how Zone V in your digital image file in Photoshop in a particular RGB color space gets printed to a density of Zone V on a specific printer-paper combination.  Take a deep breath and trust that color management is the way to go about characterizing this relationship.

Your first question should really be, what is Zone V in a device-(in?)dependent definition?  Is it L*50?  Is it the black density that your printer and ink can produce that is halfway between maximum black and paper white?  At least these definitions have a physical meaning, in terms of your destination, which I hope like hell is printed output.

You may want to go to the Piezography.com website and download the free Piezography profiles - you will have no use for them, but in that download there is a user manual that you might want to read, particularly the Grayscale Management section.

Have fun!

kirk

« Last Edit: February 14, 2018, 09:38:05 am by kirkt »
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: No color management
« Reply #63 on: February 14, 2018, 11:06:14 am »

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
Yep.

Torger's printing, measurement, and profiling breakdown is a particularly excellent exposition. Highly recommended to really wrap one's mind around color, printers, and profiling. Just an outstanding piece.

texshooter,
Please review all the comments made by the folks here. They are quite good. Really thoughtful and trying to explain things in various ways. This can be difficult when first going into the gory details. But you are getting really excellent advice. Probably because you are also doing work and have an instrument that you can use and explore with and are willing to do so.
Logged

joofa

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 544
Re: No color management
« Reply #64 on: February 14, 2018, 11:52:05 am »


RGB color spaces are artificial representations of color

How can all RGB spaces be artificial when CIE did original experiment with an RGB space?

Quote
In contrast, Profile Connection Spaces (PCS) are based in the physiology of how a "Standard" human observer perceives light of a specific "tristimulus" value - that is, how our visual systems interpret that light. 

Original CIE RGB space(s) were "based in the physiology of how a "Standard" human observer perceives light of a specific "tristimulus" value".

And, a frequently used, so called, Profile Connection Space, is XYZ. Which is just another RGB space - after all it is a matrix multiple of CIE RGB space.

Quote
These physically-based models of color form the basis for our ability to relate various artificial, device-dependent RGB models to each other - as such, they are DEVICE-INDEPENDENT.  Hence the term "profile connection spaces" -

There is nothing stopping making CIE 1931 RGB the "profile connection space" (as mentioned it is just a matrix transform away from XYZ). However, some how XYZ has come to be that canonical one.

Quote
the physically-referred PCS can connect one artificial RGB space to another, as long as the relationship between the RGB space and the PCS is known. 

IMHO, no artificiality here in general. CIE did original experiments in an RGB space. That can't be an artificial space! If anything, XYZ could be called artificial in some sense, after all its primaries are outside the visible gamut.
Logged
Joofa
http://www.djjoofa.com
Download Photoshop and After Effects plugins

kirkt

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 604
Re: No color management
« Reply #65 on: February 14, 2018, 12:29:43 pm »

I think, perhaps, this conflates R, G and B spectral representations of light with R, G and B numerical values in a device-dependent color model.  I should have been more succinct and stated that the color spaces (ProPhoto, AdobeRGB, sRGB) that the OP is referencing.... Again, I am engaging in simplification to explain the general aspects of color management in the context of the OP's query.  I am not qualified to debate the history and physics of color science.

kirk
« Last Edit: February 14, 2018, 12:34:18 pm by kirkt »
Logged

joofa

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 544
Re: No color management
« Reply #66 on: February 14, 2018, 12:33:18 pm »

I think, perhaps, this conflates R, G and B spectral representations of light with R, G and B numerical values in a device-dependent color model.  Again, it is simplification to explain the general aspects of color management in the context of the OP's query.  I am not qualified to debate the history and physics of color science.

kirk

The point is how can XYZ be 'device independent', PCS, and CIE 1931 RGB is not by your estimation. The only difference between them is a matrix transform. Thats all.
Logged
Joofa
http://www.djjoofa.com
Download Photoshop and After Effects plugins

kirkt

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 604
Re: No color management
« Reply #67 on: February 14, 2018, 12:35:45 pm »

The point is how can XYZ be 'device independent', PCS, and CIE 1931 RGB is not by your estimation. The only difference between them is a matrix transform. Thats all.

See my edit, referencing the OP's color spaces used.

kirk
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: No color management
« Reply #68 on: February 14, 2018, 01:03:43 pm »

The point is how can XYZ be 'device independent', PCS, and CIE 1931 RGB is not by your estimation. The only difference between them is a matrix transform. Thats all.

Not the only difference. Standard RGB spaces must first be "de-gammafied," put into linear space, then a matrix transform applied, then re-gammafied to the target RGB space.

Generally, RGB spaces, outside of things like ProPhoto (ROMM) RGB, are designed to approximate something physically realizable for color additive devices like monitors, projectors, and such. The XYZ PCS space is a convenient intermediary but not strictly necessary for RGB space conversions. It's major convenience is that the XYZ values are linear and positive. The CIERGB space requires negative numbers to represent many colors in the human gamut. If one allows negative numbers, any RGB colorspace, even sRGB, can represent all of the CIE gamut colors.

The LAB PCS space is normally used for subtractive color devices such as printers. It make 3D lookup tables feasible. However, the ICC LAB PCS, unlike CIELAB, fails as a general colorspace for additive RGB devices since some of them exceed the limits if ICC LAB PCS which clips a* and b* at -128.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20708
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: No color management
« Reply #69 on: February 14, 2018, 01:08:39 pm »

How can all RGB spaces be artificial when CIE did original experiment with an RGB space?
Only some like RGB working spaces which are all synthetic (including sRGB of course). Nothing artificial about the RGB ICC profiles for my Epson's.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

joofa

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 544
Re: No color management
« Reply #70 on: February 14, 2018, 01:37:32 pm »

Not the only difference. Standard RGB spaces must first be "de-gammafied," put into linear space, then a matrix transform applied, then re-gammafied to the target RGB space.

I mentioned standard XYZ and CIE 1931 RGB. I think these spaces don't have a gamma.
Logged
Joofa
http://www.djjoofa.com
Download Photoshop and After Effects plugins

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: No color management
« Reply #71 on: February 14, 2018, 01:49:29 pm »

I mentioned standard XYZ and CIE 1931 RGB. I think these spaces don't have a gamma.

They aren't, of course, but essentially all RGB spaces in common usage have a gamma or modified gamma.. I actually use a modified Adobe RGB profile with a gamma=1 in Photoshop, but only with high bit resolution. It produces image resizing without the Moire risk.

XYZ also has the advantage that it's self descriptive outside of a global intensity magnitude. Others need to have the primaries specified. Even LAB requires a reference white point (in XYZ) to define a specific color though D50 is typically just assumed.
Logged

joofa

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 544
Re: No color management
« Reply #72 on: February 14, 2018, 02:36:42 pm »

They aren't, of course, but essentially all RGB spaces in common usage have a gamma or modified gamma..

Yes. However, IMHO, I don't think that adding gamma changes the device independence or dependence characterization.

Quote
I actually use a modified Adobe RGB profile with a gamma=1 in Photoshop, but only with high bit resolution. It produces image resizing without the Moire risk.

That is interesting. I think Bartvander Wolf did some experiments on gamma / no-gamma in image resizing and reported here on LL. I have forgotten what was his prognosis.

Quote
XYZ also has the advantage that it's self descriptive outside of a global intensity magnitude. Others need to have the primaries specified. Even LAB requires a reference white point (in XYZ) to define a specific color though D50 is typically just assumed.

LAB is (mostly) glorified non-linear (gamma-imposed) XYZ color space. But, if one is digging more into color spaces, then there are at least 4 different representations of a 'color space' from the same data / experiments. They are listed here:

Color Space Representations

Usually, the colorimetry community has focused on the representation #1 only. However, others are quite useful for certain calculations. For e.g., here it used to show that a frequent misconception that daylight is green is misconstrued:

Is daylight green?

The interesting thing is that since the 4 representations in the first link are different - would yield different numerics - then what is an 'absolute' color space? One would possibly have to settle that a metameric black color space is the real absolute space. Because, one can't make metameric colors to become suddenly visually differentiated by only using a different mathematical representation.

So all color spaces are 'artificial'  ;D I guess, that could be seen as a generalization of what Kirkt said - though probably not in the way intended.
« Last Edit: February 14, 2018, 02:41:02 pm by joofa »
Logged
Joofa
http://www.djjoofa.com
Download Photoshop and After Effects plugins

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: No color management
« Reply #73 on: February 14, 2018, 04:18:32 pm »


The interesting thing is that since the 4 representations in the first link are different - would yield different numerics - then what is an 'absolute' color space? One would possibly have to settle that a metameric black color space is the real absolute space. Because, one can't make metameric colors to become suddenly visually differentiated by only using a different mathematical representation.

So all color spaces are 'artificial'  ;D I guess, that could be seen as a generalization of what Kirkt said - though probably not in the way intended.

True, Lots of ways too look at things. The reasonably bright LED flashlight in my hand emits fewer visible photons than my hand, that is holding it, emits invisible photons. Since we can't see the latter, we don't think about them.
Logged

joofa

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 544
Re: No color management
« Reply #74 on: February 14, 2018, 04:28:28 pm »

True, Lots of ways too look at things.

What is implicit in the definition of a color space, though hardly talked about, is a certain weight function that defines the mapping between radiant energy and color representation via so called 'color numbers'. More on it here:

Color Spaces, CIE and Otherwise - The Hidden Weight Matrix
Logged
Joofa
http://www.djjoofa.com
Download Photoshop and After Effects plugins

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: No color management
« Reply #75 on: February 14, 2018, 05:04:55 pm »

What is implicit in the definition of a color space, though hardly talked about, is a certain weight function that defines the mapping between radiant energy and color representation via so called 'color numbers'. More on it here:

Color Spaces, CIE and Otherwise - The Hidden Weight Matrix

Well, we benefit from using a canonical form, even with it's imperfections.  I like your point about the subset of metamers remaining metamers regardless of choice of colorspaces.

Getting back to this thread's topic, I looked at three different printer paper combos. A glossy with the Canon 9500II and Epson 9800 as well as a matte with the 9800. The native drivers (w/o color management) printed RGB 128,128,128 as L* 34, 57, and 40 respectively so there is quite a variation in device space RGB printed colors.
Logged

Stephen Ray

  • Full Member
  • ***
  • Offline Offline
  • Posts: 218
Re: No color management
« Reply #76 on: February 14, 2018, 05:37:55 pm »

Getting back to this thread's topic, I looked at three different printer paper combos. A glossy with the Canon 9500II and Epson 9800 as well as a matte with the 9800. The native drivers (w/o color management) printed RGB 128,128,128 as L* 34, 57, and 40 respectively so there is quite a variation in device space RGB printed colors.

These discrepancies are explained how?
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: No color management
« Reply #77 on: February 14, 2018, 05:53:00 pm »

These discrepancies are explained how?

Native device RGB spaces are somewhat arbitrary and that is somewhat necessary as they have to map all the RGB values to rather non-uniform gamuts. I am not surprised by this. In fact I'd be surprised if they were all quite close.

What matters is how smoothly and repeatedly they operate. Profiles take care of the rest.
Logged

Stephen Ray

  • Full Member
  • ***
  • Offline Offline
  • Posts: 218
Re: No color management
« Reply #78 on: February 14, 2018, 06:23:02 pm »

Native device RGB spaces are somewhat arbitrary and that is somewhat necessary as they have to map all the RGB values to rather non-uniform gamuts. I am not surprised by this. In fact I'd be surprised if they were all quite close.

What matters is how smoothly and repeatedly they operate. Profiles take care of the rest.

Are you suggesting printer mfrs, (Canon & Epson in your case) "calibrate" or "pre-set" a few disparate print media of theirs to the values (or thereabouts) that you've measured? Do you imagine most machines leaving the factory would measure close to your numbers for the individual media?
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: No color management
« Reply #79 on: February 14, 2018, 06:39:07 pm »

Are you suggesting printer mfrs, (Canon & Epson in your case) "calibrate" or "pre-set" a few disparate print media of theirs to the values (or thereabouts) that you've measured? Do you imagine most machines leaving the factory would measure close to your numbers for the individual media?
They don't set it to the measured values but to a range of colors specific to media and printer. My measurements are just one point on each. Printers of the same make and model and paper will vary only a little from those readings.

For the high end professional printers each printer is individually calibrated and settings stored in the printer's electronics. They may do that as well for lower end machines though it's likely the acceptable tolerances are wider. Manufacturers normally go through cycles of refinement and cost reduction. It may well be cheaper and better to calibrate each printer separately. Alternatively, tighter manufacturing tolerances may allow them to skip the step.

My main point is that there is no special reason for manufacturers to all use the same L*  RGB targets and they clearly do not.
Logged
Pages: 1 2 3 [4] 5   Go Up