Ad
Ad
Ad
Pages: [1] 2 »   Bottom of Page
Print
Author Topic: cmyk neutral grayscale values ?  (Read 6944 times)
smilem
Full Member
***
Offline Offline

Posts: 216


« on: August 25, 2013, 06:37:48 PM »
ReplyReply

Hello, I have question about cmyk grayscale values or should I say neutral grayscale because I'm looking to use all channels.

In RGB it's easy you just send a file with equal amount of R G B like 119,119,119 produces gray patch when printed equal to LAB L50 a0 b0

To produce the same in CMYK I need to use C50 M40 Y40 K27 or LAB L50 a0 b0 right?

Second question is why photoshop colorpicker or should I say color selector is wrong all the time? If I set my color to be C50 M40 Y40 K27, then LAB should be L50 a0 b0 but it's not.
LAB is something like L47 a0 b-1 the HSB is also not H0 S0.

If I set the zeros by hand in the fields I said then paint with this color when you take colorpicker and check what you painted its all wrong.

I painted LAB L0 a0 b0 PS said it's CMYK C88 M79 Y65 K93
I got LAB L1 a0 b-1 PS said it's CMYK C88 M78 Y65 K93

The same happens all the time with other combinations required for neutral gray patches too, so I had to use a plugin for color selector to over come this nonsense.

Nobody noticed this? Photoshop Math functions are very bad here, color angles in HSB is a clear evidence for this, the angle should be the same for neutral LAB colors but the angle changes all the time even if you set LAB by hand to be zeros for a and b.
« Last Edit: August 25, 2013, 06:40:30 PM by smilem » Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8636



WWW
« Reply #1 on: August 25, 2013, 06:41:07 PM »
ReplyReply

To produce the same in CMYK I need to use C50 M40 Y40 K27 or LAB L50 a0 b0 right?

Only for the CMYK profile you have loaded to provide this recipe. IOW, every CMYK device is different and will have (can have) a completely different mix of CMY&K to produce a neutral.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
MarkM
Sr. Member
****
Online Online

Posts: 257



WWW
« Reply #2 on: August 25, 2013, 07:34:07 PM »
ReplyReply

Photoshop Math functions are very bad here, color angles in HSB is a clear evidence for this, the angle should be the same for neutral LAB colors but the angle changes all the time even if you set LAB by hand to be zeros for a and b.

Hue angles is indeterminate for neutral colors—you're directly in the center of the space so all the angles are the same. Kind of like being at the north pole and asking what your longitude is. Also keep in mind HSB is not an absolute color space—it's just a very simple transformation of RGB and is device dependent.
Logged

smilem
Full Member
***
Offline Offline

Posts: 216


« Reply #3 on: August 26, 2013, 02:47:46 PM »
ReplyReply

Only for the CMYK profile you have loaded to provide this recipe. IOW, every CMYK device is different and will have (can have) a completely different mix of CMY&K to produce a neutral.

I'm talking about device independent color spaces such as isocoated, gracol2006, swop etc.
The correct CMYK recipe is a ICC profile job.

But I must send correct color request to get ICC profile make grayscale, so I'm asking does every device independent color space listed above requires a different CMYK values to request printer profile to make grayscale ?
Logged
Czornyj
Sr. Member
****
Offline Offline

Posts: 1371



WWW
« Reply #4 on: August 26, 2013, 03:35:50 PM »
ReplyReply

These are CMYK color spaces of an offset press "device", so they're device depended by definition Wink

There are countless CMYK values that will build a neutral grayscale - depending on GCR/UCR curves used while creation of CMYK ICC profile.

I'm talking about device independent color spaces such as isocoated, gracol2006, swop etc.
The correct CMYK recipe is a ICC profile job.

But I must send correct color request to get ICC profile make grayscale, so I'm asking does every device independent color space listed above requires a different CMYK values to request printer profile to make grayscale ?
« Last Edit: August 26, 2013, 03:41:06 PM by Czornyj » Logged

smilem
Full Member
***
Offline Offline

Posts: 216


« Reply #5 on: August 27, 2013, 05:17:17 AM »
ReplyReply

So to create say a target for RGB grayscale you just use patches with equal amounts of RGB. And then let printer firmware/chip combo to do the conversion.

How do I proceed with the same task for native CMYK printer that mat use ISOcoated, SWOP, gracol etc?
Logged
Czornyj
Sr. Member
****
Offline Offline

Posts: 1371



WWW
« Reply #6 on: August 27, 2013, 07:26:55 AM »
ReplyReply

1) You leave the content in RGB and let the RIP do the conversion

2) You prepare the content in RGB and convert it to ISO Coated v2 (etc.) in a final step - many CMYK printers (like solvent printers, digital presses) automatically simulate ISO Coated v2 (etc.)

Personally I prefer the first workflow - ISO Coated v2 has smaller gamut than most digital presses, solvent/pigment printers, so you can get richer colours when converting RGB content directly to native CMYK profile of such device (it's good to have such profile to soft proof the effect). The second workflow may be useful when you prepare the content for an offset press, or want to maintain colours from offset in prints realised in other printing techniques
« Last Edit: August 27, 2013, 07:35:21 AM by Czornyj » Logged

MarkM
Sr. Member
****
Online Online

Posts: 257



WWW
« Reply #7 on: August 27, 2013, 03:27:21 PM »
ReplyReply

The fact that you can get neutral gray colors with all channels equal in RGB is an artifact of the way RGB working spaces are designed. They are mathematical constructs that are designed specifically to be well-behaved in this way. It's not necessarily a quality of RGB devices—if you look at RGB printer profiles and even some monitor profiles you'll sometimes find that to get grey involves sending unequal RGB values. In general monitors get a lot of this straightened out by calibrating the LUTs and curves on the card so may not notice it by the time you are profiling.

Since CMYK involves the real world with all it's messiness we can't control things as well. Imagine someone handing you four colors of paint and telling you to mix a perfectly neutral gray. One strategy may be to take equal amounts of each paint, but the odds of that working are astronomically small. A better strategy would be to have some sort of lookup table where you can specify the color you want in a more abstract color space and then use the table to find the right recipe for the real world. This is pretty much what profiles and color management do for you.
Logged

digitaldog
Sr. Member
****
Offline Offline

Posts: 8636



WWW
« Reply #8 on: August 27, 2013, 04:36:02 PM »
ReplyReply

How do I proceed with the same task for native CMYK printer that mat use ISOcoated, SWOP, gracol etc?

You must have the ICC profile that defines that process and hope it's accurately describing neutral gray balance.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
Tim Lookingbill
Sr. Member
****
Offline Offline

Posts: 1147



WWW
« Reply #9 on: August 27, 2013, 10:15:51 PM »
ReplyReply

I have a prepress "Color Process Manual" copyright 2000, used by graphic designers and bought off Amazon that's sheetfed printed. I created a 128RGB gray patch in ProPhotoRGB and noted PS's Info Palette CMYK readings converting this patch to US Sheetfed Coated v2 profile and wrote the numbers down as C40%,M30%,Y30%,K1%.

I located the CMYK combo patch of this gray on my "Color Process Manual" and isolated it with the provided crop viewer and took a picture of it lit under my 5000K daylight flotubes. See below.

I've also included a picture of the CMYK ink purity ramp between two completely different printing companies using their own ink formulation, commercial sheetfed press and coated paper. The Ole No Moire color target was printed on a Komori 300LPI press at Copy Craft in Odessa Texas and the "Color Process Manual" was printed in Hong Kong.

I know the Copy Craft color target was converted to regular US SWOP in PS5 as indicated by the company's online instructions on how to prep files for their presses.

The "Color Process Manual" 24,000 CMYK combo color swatches were printed on a Heidelberg Speedmaster w/direct to plate technology using Japan ink & glossy coated paper with these additional specs...

10% dot gain, 175LPI...

Ink densities: Black 1.8-2.0, Cyan 1.6-1.7, Magenta 1.4-1.5, Yellow 1.3-1.4 AND printed in that order.

From the manual's Photoshop 5 Color Picker screengrabs I'm assuming they converted to default PS 5 US SWOP.

It doesn't seem exact CMYK grayscale numbers are all that necessary. Derive your own conclusions.
« Last Edit: August 27, 2013, 10:23:35 PM by Tim Lookingbill » Logged
MarkM
Sr. Member
****
Online Online

Posts: 257



WWW
« Reply #10 on: August 27, 2013, 10:21:41 PM »
ReplyReply

It doesn't seem exact CMYK grayscale numbers isn't all that necessary. Derive your own conclusions.

The double negative leaves me a little unsure about your conclusion. Does that mean exact numbers are necessary?
Logged

Tim Lookingbill
Sr. Member
****
Offline Offline

Posts: 1147



WWW
« Reply #11 on: August 27, 2013, 10:29:18 PM »
ReplyReply

The double negative leaves me a little unsure about your conclusion. Does that mean exact numbers are necessary?

I caught it and changed it, Mark, before you posted.

Anything else you found valuable to comment on my contribution here?
Logged
MarkM
Sr. Member
****
Online Online

Posts: 257



WWW
« Reply #12 on: August 28, 2013, 01:07:23 AM »
ReplyReply

Anything else you found valuable to comment on my contribution here?

Tim, I'm honestly not sure what I'm supposed to take away from all this. You've taken photos of various grey swatches printed on different devices. Maybe you can be a bit more explicit about what you are demonstrating (it's also possible I'm just being a little dense and could just use a little more explanation).
Logged

Czornyj
Sr. Member
****
Offline Offline

Posts: 1371



WWW
« Reply #13 on: August 28, 2013, 02:45:41 AM »
ReplyReply

It doesn't seem exact CMYK grayscale numbers are all that necessary. Derive your own conclusions.

40/30/30 will give (an almost) neutral result when printed with european, US or japanese standarized offset press on type 1 or 2 paper - you can check it in Photoshop in a few seconds, the whole experiment wasn't all that necessary. You can also get virtually the same result with less CMY and more K combination like 35/25/25/10, or pure K 45.

Still the conclusion is that it's easier to work in RGB and let Photoshop find the magic combination of CMYK values for a given neutral shade.
« Last Edit: August 28, 2013, 02:54:09 AM by Czornyj » Logged

smilem
Full Member
***
Offline Offline

Posts: 216


« Reply #14 on: August 28, 2013, 05:49:53 AM »
ReplyReply

The commercial associations that backing up the printing standards should then provide the neutral gray recipes for their working spaces like SWOP, Gracol, ECI etc. I'm not talking about gray with only K channel as that is self explanatory and good gradations is a problem when using K only in photographic print.

I think the most precise way to convert to CMYK is by using photoshop convert to profile not just use colorpicker as that does not use any profiles for converting you color data.

The correct CMYK ink values on device side is managed by ICC profile for that device and I don't care how it's doing the magic to make neutrak gray, I however do care that the color recipe sent to the printer would be correct else the gray will be bad.

In other words lets suppose I want to test certain printers grayscale and I'm making like 20 patch gray test target.
The target must be made for 3 standards like SWOP, Gracol, ECI.
I convert the neutral gray patches from RGB to the profiles above in Photoshop.

I see that the CMYK recipes are all over the place they not very close to each other and then when I look at LAB values I see they thay are not even gray. If I enter correct lab values for a and b the color changes to more gray looking then the Photoshop was able to make by using the ICC profile conversion.

So in the end I still do not understand what to use....    as LAB is used in every case from RGB to CMYK then I should correct the CMYK values by altering LAB in Photoshop after conversion to SWOP, Gracol, ECI right?

Then if I feed this to a CMYK printer that opens the file as SWOP, Gracol, ECI colorspace with neutral gray it feeds the data to internal CMM/RIP to do the conversion to actual CMYK values to put on paper by profile for that paper.
« Last Edit: August 28, 2013, 06:04:11 AM by smilem » Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8636



WWW
« Reply #15 on: August 28, 2013, 06:17:31 AM »
ReplyReply

I see that the CMYK recipes are all over the place they not very close to each other and then when I look at LAB values I see they thay are not even gray. If I enter correct lab values for a and b the color changes to more gray looking then the Photoshop was able to make by using the ICC profile conversion.

So in the end I still do not understand what to use....    as LAB is used in every case from RGB to CMYK then I should correct the CMYK values by altering LAB in Photoshop after conversion to SWOP, Gracol, ECI right?

Lab is device independent. CMYK isn't even close (just the opposite). RGB working spaces are Quasi-Device Independent. IOW, RGB isn't device independent but because the RGB working spaces are theoretical constructs. What you want (CMYK that in call cases has set's of values that are neutral) require a good ICC profile for the specific process that will convert RGB and send CMYK to that device.
« Last Edit: August 28, 2013, 06:27:35 AM by digitaldog » Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
smilem
Full Member
***
Offline Offline

Posts: 216


« Reply #16 on: August 28, 2013, 06:34:01 AM »
ReplyReply

So what happens if I send LAB to CMYK printer? The printer profile avoids the RGB conversion to LAB side by making the conversion more precise?

I agree that to test a RGB->CMYK devicelink profile it's best to send RGB data.
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8636



WWW
« Reply #17 on: August 28, 2013, 06:38:28 AM »
ReplyReply

So what happens if I send LAB to CMYK printer? The printer profile avoids the RGB conversion to LAB side by making the conversion more precise?
I agree that to test a RGB->CMYK devicelink profile it's best to send RGB data.

A device link is no different here. It's a recipe like the ICC profile. If it's right, you get what you expect (in this context a neutral). If not, you don't. You can send RGB and the next person in the workflow either has or hasn't got the right recipe. There is no such thing as a Lab printer. A conversion still has to take place if you send Lab or tagged RGB. Lab doesn’t solve this problem one bit.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
smilem
Full Member
***
Offline Offline

Posts: 216


« Reply #18 on: August 28, 2013, 09:34:56 AM »
ReplyReply

So to create say a target that is true gray I have to select proper working space like SWOP, Gracol, ECI.

Then I have to create the patches from scratch by entering LAB values and let Photoshop calculate proper CMYK right?
And I have to do this for every working space all over again.

So the CMYK recipe calculated by photoshop is the best or there are other options? Like I tried the Color Composer plugin and IMO it's better, more easy to use, math seems to wander less. But it's old and I don't expect 64bit version anytime soon.
« Last Edit: August 28, 2013, 09:38:06 AM by smilem » Logged
MarkM
Sr. Member
****
Online Online

Posts: 257



WWW
« Reply #19 on: August 28, 2013, 12:42:00 PM »
ReplyReply

I think the most precise way to convert to CMYK is by using photoshop convert to profile not just use colorpicker as that does not use any profiles for converting you color data.

The color picker does use a profile. It's the CMYK profile set in photoshop's color settings.

It might help if you tell us what you are trying to accomplish.  I can't think of any reason why you would want to send a bunch of different grays from different spaces to a printer. If you just want to be sure you are getting neutral gray, a better solution is to talk to the printer and ask what profile to use. Certainly quicker and cheaper than guess & check. Then if you have a problem with a proof you can say, look, your profile says this should be neutral. This is what most people do, and it works pretty well (if you have a good printer—but that's a whole different can of worms).

If you really want to dig into the specifications (probably a good idea before you start criticizing them), a good place to start might be IdeAlliance: http://www.idealliance.org/specifications/g7 but be warned, it's aimed at printing professionals, not photographers. This is a good overview of the problem of neutral colors in commercial printing: http://www.npes.org/Portals/0/standards/pdf/ANSI-CGATS-TR015-2013.pdf
Logged

Pages: [1] 2 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad