Colour Printing Question

[William Walker]

I have been asked to print a chart with specific CMYK values supplied to me with a U.S. web coated (swop) v2 profile.

I made the print out of Photoshop - via the Canon Print Tool - without altering the colour-space, but using my own custom profile for the paper.

Looking at the print in the viewing booth, as I do with all my prints, the colours seem accurate to my eye (and the customer's satisfaction). However, when I measure the values on the print using ColoMunki Spot Measurement, the numbers are, in most cases, not the same as the originals, some worse than others as can be seen on the attached image.

I have a few questions:
1)  Is it my paper profile that is not accurate enough?
2)  Is it unrealistic to expect the numbers to be 100% accurate in the first place?
3)  Should I be converting the profile to RGB?
4)  Will rendering intent be a factor? If so, which would be the best to use?

In the attached image I have left the desired CMYK colour (numbers at the bottom) on the left-side of the square. The right-hand side of the square has the numbers and colours from the print.

I am interested to know what I am doing wrong!

Thanks,
William

[digitaldog]

You're not doing anything wrong. The measurement of the print may differ from the ”original” which is a different print process right?

[William Walker]

You're not doing anything wrong. The measurement of the print may differ from the ”original” which is a different print process right?

Does that mean my second question is relevant... I am expecting too much?

[GWGill]

Does that mean my second question is relevant... I am expecting too much?

It's not clear from your post whether you are throwing CMYK values at your printer, or whether you are converting the colors for your particular printer (i.e. you are using the SWOP profile as a source, and your printer profile as a destination). If it's the former, then it would be surprising if the colors matched expectation. If it is the latter, then you expectations should be higher, but the accuracy will depend on whether colors are in gamut, whether your rendering intent matches your color measurement technique, and whether your printer profile is accurate. If you are visually comparing the SWOP swatches, then there is the factor of whether the SWOP profile matches the swatches.

You aren't giving us any objective information though (i.e. delta E's) - what are the actual color measurements vs. the SWOP profile prediction ?
Are you measuring and comparing white point relative, or absolute color values ?

[William Walker]

It's not clear from your post whether you are throwing CMYK values at your printer, or whether you are converting the colors for your particular printer (i.e. you are using the SWOP profile as a source, and your printer profile as a destination). If it's the former, then it would be surprising if the colors matched expectation. If it is the latter, then you expectations should be higher, but the accuracy will depend on whether colors are in gamut, whether your rendering intent matches your color measurement technique, and whether your printer profile is accurate. If you are visually comparing the SWOP swatches, then there is the factor of whether the SWOP profile matches the swatches.

You aren't giving us any objective information though (i.e. delta E's) - what are the actual color measurements vs. the SWOP profile prediction ?
Are you measuring and comparing white point relative, or absolute color values ?

Hi,

Thanks for your reply. I will try and answer your questions (although I'm not sure I understand them!! )

I am not converting the CMYK. I simply export (as I receive it) to the Canon Print Plug-in using the correct profile for the paper. Perceptual Rendering.

As far as delta E's is concerned; are the readings on the attachment not what you are looking for?  The "desired" CMYK (Swop prediction??) values (as per Photoshop) are at the bottom of each square with their corresponding colours on the left-hand half of the square. The number and the values on the right-hand half of the square are the "actual" readings I took off the print. In simple terms, the left-hand side of each square is what I sent to the printer and the right-hand side are the colours I measured off the print.

I am afraid I did not understand much else! I will not be offended in any way if you throw in the towel at this point! I really appreciate your input and I will need to looking learn more about the points you raised.
Thanks you!
William.

[rasworth]

I will try to simplify, although warning up front, I'm not on the same level of expertise as GWGill or digitaldog.  Printing with perceptual (or relative) rendering intent maps (i..e., modifies) your colors from the white point of the image space to the white point of the paper, per information stored in the profile.  Your only chance for colorimetric matching is to use absolute rendering intent.  I'm not sure what the Canon software does with CMYK inputs, this may also complicate the process, as most of the drivers are written for RGB inputs.

Note that using absolute rendering intent may "dull" the image, as the white of the paper has to be grayed somewhat to bring it in line with the image source profile white point.  If you and your customer are pleased with the image appearance using perceptual or relative intents, then you may want to stop there, as this is the normal printing method.  The eye tends to "see" the colors in relation to the surrounding white, so that you may achieve a superior visual match using perceptual or relative, as long as you don't do a side-by-side compare.

BTW you may obtain a better match using Relative intent, Perceptual throws in more color changes, trying in various ways to "squeeze" more of the source colors into the limited range of the printer/paper combination.

Richard Southworth

[William Walker]

I will try to simplify, although warning up front, I'm not on the same level of expertise as GWGill or digitaldog.  Printing with perceptual (or relative) rendering intent maps (i..e., modifies) your colors from the white point of the image space to the white point of the paper, per information stored in the profile.  Your only chance for colorimetric matching is to use absolute rendering intent.  I'm not sure what the Canon software does with CMYK inputs, this may also complicate the process, as most of the drivers are written for RGB inputs.

Note that using absolute rendering intent may "dull" the image, as the white of the paper has to be grayed somewhat to bring it in line with the image source profile white point.  If you and your customer are pleased with the image appearance using perceptual or relative intents, then you may want to stop there, as this is the normal printing method.  The eye tends to "see" the colors in relation to the surrounding white, so that you may achieve a superior visual match using perceptual or relative, as long as you don't do a side-by-side compare.

BTW you may obtain a better match using Relative intent, Perceptual throws in more color changes, trying in various ways to "squeeze" more of the source colors into the limited range of the printer/paper combination.

Richard Southworth

Hi Richard,

Thanks for your reply.

After reading and then looking up a few things from GWGill, I created a new chart from scratch in PS in ProPhoto using the desired CMYK numbers in the "Fill" - "Colour" process.

I then made a print using "Absolute Colormetric", working on the idea that CMYK colours would be within the gamut of the profile for the specific paper. The results are damn near...!

I am quite satisfied with the result which is more than good enough for the printing I do.

Thank you for taking the time to assist!

Regards

William

[Alistair]

As a matter if interest, for the colour which achieved the closest match; is the spectral data for each colour very close ?

[William Walker]

Hi Alistair,

Yes, they are.

[Alistair]

Interesting. How many colours does your printer have? Is there any metameric failure as you view the comparable swatches under different light sources? (other than the presumably D50 of your measuring device?). Spectral curve matching of spot colours is the holy grail for me but by goodness it is difficult!

[William Walker]

Interesting. How many colours does your printer have? Is there any metameric failure as you view the comparable swatches under different light sources? (other than the presumably D50 of your measuring device?). Spectral curve matching of spot colours is the holy grail for me but by goodness it is difficult!

Hi Alistair,

It is a Canon iPF8400 which has 12 inks.

As. for your other questions, I simply measured the values in Photoshop and then measured the prints with the Colormunki Photo and compared those numbers and as per the visual closeness the numbers were close too.

I do not claim to be an expert in this area, it was a passing interest. I print photography and am more than satisfied with the results...I got down into the weeds just this once!

[Alistair]

I understand, no problem. Thanks.

[aaron125]

Hi Richard,

Thanks for your reply.

After reading and then looking up a few things from GWGill, I created a new chart from scratch in PS in ProPhoto using the desired CMYK numbers in the "Fill" - "Colour" process.

I then made a print using "Absolute Colormetric", working on the idea that CMYK colours would be within the gamut of the profile for the specific paper. The results are damn near...!

I am quite satisfied with the result which is more than good enough for the printing I do.

Thank you for taking the time to assist!

Regards

William

When you mention “desired CMYK numbers”, what does that mean, exactly? On their own, as in, without a CMYK profile to go with your CMYK numbers, are literally worthless. They are worthless by definition, in that they literally have no (device independent/Lab) values associated with them. The easiest way to picture this is to think of a large TV shop which has myriad TVs all playing the same thing, yet each individual TV is displaying the same colours slightly differently. Reason being is that they are all receiving (for example) RGB numbers but, just like your CKYK numbers, the RGB numbers are similarly worthless. If an LG TV used slightly differently manufactured LEDs then their colour balance obviously won’t be identical. So even if both sets are displaying R100,G46,B37, or any other RGB value, because they use physically different LED backlights and LCD panels (and of course, different components in the various circuits), there’s no way for them to display identical colours.

If they were displaying sRGB 100,46,37 then the 2 TVs just might have a chance of displaying identical output. They’d need to be calibrated and profiled but at least now the colours are being specified unambiguously and so there’s at least a slight chance of it being possible. It’s the profile which makes the RGB/CMYK numbers unambiguous.

So, you need to specify your “desired CMYK numbers” either with a profile or better still would be to just specify the colours you want in Lab as it is a device independent colour space and there aren’t any variants of Lab, there’s just one Lab space.

I realise from your previous replies that you don’t have a thorough knowledge and understanding of colour management. Without at least a working understanding of the basics of colour management, you will find your request incredibly frustrating and as you already said, you don’t understand what you’re doing wrong. My best advice is to read up on colour management, get a basic understanding of how the main concepts work and you’ll be on your way. Good luck with it.👍

PS: from where or how did you obtain the CMYK numbers you desire?

[William Walker]

When you mention “desired CMYK numbers”, what does that mean, exactly? On their own, as in, without a CMYK profile to go with your CMYK numbers, are literally worthless. They are worthless by definition, in that they literally have no (device independent/Lab) values associated with them. The easiest way to picture this is to think of a large TV shop which has myriad TVs all playing the same thing, yet each individual TV is displaying the same colours slightly differently. Reason being is that they are all receiving (for example) RGB numbers but, just like your CKYK numbers, the RGB numbers are similarly worthless. If an LG TV used slightly differently manufactured LEDs then their colour balance obviously won’t be identical. So even if both sets are displaying R100,G46,B37, or any other RGB value, because they use physically different LED backlights and LCD panels (and of course, different components in the various circuits), there’s no way for them to display identical colours.

If they were displaying sRGB 100,46,37 then the 2 TVs just might have a chance of displaying identical output. They’d need to be calibrated and profiled but at least now the colours are being specified unambiguously and so there’s at least a slight chance of it being possible. It’s the profile which makes the RGB/CMYK numbers unambiguous.

So, you need to specify your “desired CMYK numbers” either with a profile or better still would be to just specify the colours you want in Lab as it is a device independent colour space and there aren’t any variants of Lab, there’s just one Lab space.

I realise from your previous replies that you don’t have a thorough knowledge and understanding of colour management. Without at least a working understanding of the basics of colour management, you will find your request incredibly frustrating and as you already said, you don’t understand what you’re doing wrong. My best advice is to read up on colour management, get a basic understanding of how the main concepts work and you’ll be on your way. Good luck with it.👍

PS: from where or how did you obtain the CMYK numbers you desire?

Thanks! 👍