Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: Conventional CMYK  (Read 5318 times)
opgr
Sr. Member
****
Offline Offline

Posts: 1125


WWW
« on: May 11, 2006, 10:31:01 AM »
ReplyReply

Following is a small uhh..., tutorial if you will, which, just for fun, shows a conventional RGB to CMYK conversion in the digital realm. It is based on the idea that inverting RGB produces CMY.

1. Convert To Profile: ColorMatch RGB
First we have to bring the image into a more reasonable colorspace. In the early days, the RGB to CMY relation was fixed in the machines available. Later display RGB was assumed. It was certainly not something as big as AdobeRGB or ProphotoRGB.
So there must be some sort of relation between the RGB primaries and the CMYK colorspace. ColorMatch fits the bill, which is a slightly adjusted sRGB to better match the SWOP space.

2. Add a graywedge at the bottom of your image
Next add a graywedge to the image. This will enable us to rebalance the graybalance after conversion, because whereas equal amounts of RGB produce pure gray, equal amounts of CMY does not.



Now Using the Channels Palette:
3. Add an empty channel
4. Select Split Channels
5. Select Merge Channels

Because of the added channel we can now select mode CMYK and select OK.

And because of Photoshop's internal represention of channel values in CMYK, which are the inverse of channel values in RGB, we automatically end up with an inverted RGB image (and an additional empty black channel.)

The image will likely be designated as Untagged CMYK, which means Photoshop uses the CMYK profile from the colorpreferences. If this isn't set to SWOP then make sure to select Assign Profile and assign the SWOP v2 profile.

At this point you should see a very pale and muddy version of your original image.



The image looks muddy because equal amounts of CMY does not produce pure gray. The image is not colorbalanced, more or less equivalent to having an improper whitebalance set in RAW conversion. The image looks pale because C=M=Y=100% is neither gray, nor dark.

6. GrayBalance
To colorbalance the image we can use the graywedge from step 1 and the curves dialog. Open the curves dialog (as opposed to the levels dialog), and select the middle gray pipet. Next select the middle of the graywedge, and sample to let photoshop automatically correct middle gray to pure gray.




7. Increase Color
Next we have to increase the saturation slightly. CMYK does not act linearly as RGB does which is why you don't end up with pure gray above, nor pure black for maximum CMY, and also results in saturation differences. This would normally be part of the internals of the conversion engine. Select Hue/Saturation and increase saturation to taste, in this case +40 is used.




8. Black Generation
The image will still look pale because the darkest color C=M=Y=100% simply doesn't come close to reasonable dark. So now we will introduce a quick & dirty black plate:

9. Select Image->Adjust->Selective Color
10. Set the following values:
C = -35%
M = -50%
Y = -50%
K = +100%
Method = Absolute

This tells the Selective Color command to adjust C=M=Y=100% to a more reasonable under color gray of:
C = 65%,
M = 50%,
Y = 50%
And add 100% black as a substitute.



Selective Color automatically decreases its effect the further away from pure black we get. So 50% gray is hardly affected, as are mildly dark colors. What Photoshop CMYK separation settings are equivalent to this black generation is left as an exercise to the reader.



So, there you have it, a mighty fine CMYK separation. Does it bare any relation to the original? Well, if it does, fine, if it doesn't, tough luck. Make 10 proofs, adjust the freaking heck out of it, and you will have an idea what conventional prepress was all about... (read Andrew's article for a more rational explanation).
Logged

Regards,
Oscar Rysdyk
theimagingfactory
digitaldog
Sr. Member
****
Offline Offline

Posts: 9225



WWW
« Reply #1 on: May 11, 2006, 02:50:12 PM »
ReplyReply

The best quote I've heard on the subject (from author David Blanter):

God created RGB. Man Created CMYK. What would you rather use?
Logged

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

Posts: 1125


WWW
« Reply #2 on: May 14, 2006, 06:53:50 AM »
ReplyReply

Quote
The best quote I've heard on the subject (from author David Blanter):

God created RGB. Man Created CMYK. What would you rather use?
[a href=\"index.php?act=findpost&pid=65123\"][{POST_SNAPBACK}][/a]
Yeah, I read that. It's a funny quote. However, by the same token; God created bird, man created F16, which apparently leads some people to believe that only a highly trained, highly skilled individual is able to make CMYK fly...  

And actually, I'm not even opposed to that idea. An adequate and relevant CMYK conversion does take knowledge and skill. That doesn't mean that this skill can not be automated. As a matter of fact, given the decreasing number of people that really understand the trade (by experience), it is more like a prerogative or a mandate to retain this knowledge by automation. Implementing the skill (read controls) in software using defaults (read experience) to build open standard conversions in ICC profiles.

ICC profiles then become more like a great vehicle for retaining this knowledge. Just another way of looking at it, I suppose.
Logged

Regards,
Oscar Rysdyk
theimagingfactory
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #3 on: May 14, 2006, 12:45:19 PM »
ReplyReply

Okay, I obviously missed something: Why was the thread regarding Nigels article erased? There was plenty of information in that thread. Why wasn't it just closed? I find it extremely annoying to come back here, and see this thread and my postings and answers to them just gone with the wind, without any comment.

Sorry for beeing OT.

Dennis.
Logged

Best Regards

Dennis.
bruce fraser
Jr. Member
**
Offline Offline

Posts: 51


« Reply #4 on: May 14, 2006, 01:45:02 PM »
ReplyReply

Quote
Okay, I obviously missed something: Why was the thread regarding Nigels article erased? There was plenty of information in that thread. Why wasn't it just closed? I find it extremely annoying to come back here, and see this thread and my postings and answers to them just gone with the wind, without any comment.

Sorry for beeing OT.

Dennis.
[a href=\"index.php?act=findpost&pid=65432\"][{POST_SNAPBACK}][/a]

See Michael's message re "Hacker Attack" in the Digital Image Processing forum...
Logged
jlmwyo
Jr. Member
**
Offline Offline

Posts: 78


« Reply #5 on: May 14, 2006, 01:55:27 PM »
ReplyReply

Quote
Okay, I obviously missed something: Why was the thread regarding Nigels article erased? There was plenty of information in that thread. Why wasn't it just closed? I find it extremely annoying to come back here, and see this thread and my postings and answers to them just gone with the wind, without any comment.

Sorry for beeing OT.

Dennis.
[a href=\"index.php?act=findpost&pid=65432\"][{POST_SNAPBACK}][/a]

Ugg, there was a LOT of good info in that thread. Sad.....
Logged

Images of Wyoming

Link To My Gallery

Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #6 on: May 14, 2006, 03:16:23 PM »
ReplyReply

Quote
See Michael's message re "Hacker Attack" in the Digital Image Processing forum...
[a href=\"index.php?act=findpost&pid=65437\"][{POST_SNAPBACK}][/a]
Thanks for clarifying. Strange though.
Logged

Best Regards

Dennis.
macintosh
Newbie
*
Offline Offline

Posts: 21


WWW
« Reply #7 on: May 16, 2006, 02:14:07 AM »
ReplyReply

Seems a bit strange to me.......must have threatened someone......Romania my arse.....


Nigel



p.s. anyone who wishes to continue with the discussion then by all means nigelkwilliams@hotmail.com

But I will continue to investigage the possibilities of an alternative method once I have read the book that was recommended.

Thankyou opgr.

so profiles now rule? well not for everyone.........

I will get back and create a thread specifically for this topic, in the mean time many many thanks to all who helped.

Kind regards

Nigel
Logged
macintosh
Newbie
*
Offline Offline

Posts: 21


WWW
« Reply #8 on: May 16, 2006, 02:40:09 AM »
ReplyReply

so where were we?

Discussing the standardisation of colour to press and the elimination of customised ICC profiles to single standards.

Then perhaps we can remove the complications being experienced throughout our industry.

Alternatively the method of separation I detailed in the orginal article allows one to know the Photoshop colour separation processes specifically for chosen presses.

This whole argument regarding colour calibration has been waging since the early ninties and now 15 years later we are nowhere near feeling comfortable about inhouse separations and proofing. It was clearly stated then "standard ICC profiles must be developed and utilised on both sides".

So lets have an open discussion that may take time and effort, one that makes it a cheaper, easier and less complicated solution to exactly separate colours in the CMYK space for a proofer or printing press, other than the ICC profile cowboys riding into town and holding you to ransom.

rgds

Nigel
« Last Edit: May 16, 2006, 03:29:47 AM by macintosh » Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 9225



WWW
« Reply #9 on: May 16, 2006, 07:48:59 AM »
ReplyReply

Quote from: macintosh,May 16 2006, 01:40 AM
so where were we?
Discussing the standardisation of colour to press and the elimination of customised ICC profiles to single standards.
Quote

Impossible as was already mentioned. At least for all CMYK processes. Why not stick with standards for SWOP and Sheetfed (and some would say newsprint).

Then perhaps we can remove the complications being experienced throughout our industry.
Quote

Who'd industry? The print industry just needs to aim for standards in output behavior and define them (using the method most modern software products use cross platform; ICC profiles) and keep their devices in calibration.

Quote
Alternatively the method of separation I detailed in the orginal article allows one to know the Photoshop colour separation processes specifically for chosen presses.

Maybe for your shop. What about everyone else?

Quote
This whole argument regarding colour calibration has been waging since the early ninties and now 15 years later we are nowhere near feeling comfortable about inhouse separations and proofing. It was clearly stated then "standard ICC profiles must be developed and utilised on both sides".

Standard ICC profiles only define standard output behavior.

The printer I discussed in the dead thread (Basin Printing in Durango CO) DOES have both their presses and proofer behaving in a standard behavior defined by SWOP TR001. The canned U.S. Web Coated (SWOP) v2 separation as well as the custom profile I made behave roughly the same and produce excellent color BECASUE Basin conforms to a standard and keeps their process consistent to that behavior. Any photographer can build beautiful separations for Basin (and they can request and use Basin's home built profile or use the U.S. Web Coated (SWOP) v2 profile).
Logged

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

Posts: 82


« Reply #10 on: May 16, 2006, 10:29:29 AM »
ReplyReply

Quote
So lets have an open discussion that may take time and effort, one that makes it a cheaper, easier and less complicated solution to exactly separate colours in the CMYK space for a proofer or printing press, other than the ICC profile cowboys riding into town and holding you to ransom.
Well, do I misunderstand you, or do you have some personal dislikes against ICC profiles?

IMHO emotions do not really help in this issue. I don't know what could be easier than working on an image in RGB - thus keeping all the features of Photoshop - with the soft proof activated (at least from time to time), and then finally convert it either to a specific ICC profile of a certain press or to a standard profile (like SWOP TR001) and sending it to a press, which is matched to this standard?

I am doing prepress work for some years now, as part of my job. I am surely no CM guru, nor do I have as much experience as you. But in the erased discussion, it seemed to me, that you were stuck in some niche down under, working tightly withone or another press. You seem to match your Photoshop to the press 'the hard way' by turning all the small knobs at separation type, black genaration, dot gain, ink limits etc. Of course, with years of experience you get brilliant results.

But why should one go your way, if he can go the easy way? It's like not using layer styles, because you can add an embossing effect manually.

Dennis.
Logged

Best Regards

Dennis.
bruce fraser
Jr. Member
**
Offline Offline

Posts: 51


« Reply #11 on: May 16, 2006, 12:02:49 PM »
ReplyReply

I'll try to keep this very simple.

Nigel, you're confusing several different kinds of 'standards'.

A 'standard' ICC profile is simply a profile formatted according to the ICC specifications. It has recently been approved as an International Standard, ISO 15076. You can get the profile spec from the ICC's home page, www.color.org. There is no implication that it has to describe a "standard" device behavior.

There are 'standards' for print manufacturing—SNAP, GraCOL, SWOP, FOGRA, etc. If you want to evangelize the print manufacturing industry to print to these (or other) standards, I'm all for it, but it's very difficult to convince small-to-medium-size high-end sheetfed shops that they should deliberately make their press behave less than optimally just to meet a standard. Were you to succeed in this, one of the benefits would be that print buyers could use standard ICC profiles that describe the manufacturing standard, such as Adobe's US Web Coated (SWOP) v2.0.

But if the print house doesn't print to a standard, a profile that describes the standard isn't useful. You need a profile (or more likely, a family of profiles with different KGen to satisfy different image types) that describes the actual print conditions to make optimal separations for those print conditions.

There is no 'standard' behavior for displays or for capture devices (scanners and digital cameras), nor is there likely to be one in our lifetimes, because the physical capabilities of the devices keep improving. Adobe has provided 'standard' RGB editing spaces and has taken the monitor out of the loop by making the display a side-branch in the color reproduction chain, but you still need accurate custom (but still 'standard') ICC profiles that describe the behavior of your display (to get it to display colors correctly), and your capture devices (to get the captured color into a standard working space accurately).

I'm not sure who the "ICC cowboys" are that you referred to, but I'll simply observe that ICC profiling is a fifteen-year-old technology that has been mature for at least seven years, and is well-supported by companies with long track records.

More importantly, it's a technology that works. You can dick around with Custom CMYK in Photoshop until the cows come home, but it's real unlikely that the inks used on the press will come close to matching the ones described in the vintage 1992 ink models used by Custom CMYK, and a single dot gain value (even per ink) doesn't come close to the accuracy with which a good ICC profile describes the Tone Value Increase (of which dot gain is a part) that the four inks undergo when applied to the sheet. What Custom CMYK does give you is a largely illusory sense of control that ICC profiles render unneccessary.

Standard CMYK device behavior on press is possible, though not particularly likely at this point in the USA (in northern Europe it's the rule rather than the exception, so one answer is to just print everything in Germany). Standard RGB device behavior is not at all likely, since vendors compete to improve the capabilities of RGB devices. Nikon isn't about to make their cameras' behaviors match Canon's, or vice versa. Epson's $300 flatbeds probably aren't going to behave like my Imacon 848 any time soon. I don't really want my NEC 2180WG display to match a CRT from the 1990s. So standard RGB device behaviors simply isn't on the cards, nor should it be. Standard RGB editing spaces, however, are a good idea. We already have them, and the profies that describe them.

The notion that ICC profiles compensate for device behavior is fallacious. They simply describe what the device does. There's no 'correct' or 'incorrect' behavior for a scanner or a display, and while you can take a standard like SWOP as your aim point, can you say a press is behaving incorrectly if it produces a wider dynamic range or larger color gamut than SWOP?
Logged
opgr
Sr. Member
****
Offline Offline

Posts: 1125


WWW
« Reply #12 on: May 17, 2006, 03:05:01 AM »
ReplyReply

I would like to add:

The CIE initially attempted to find a reference RGB which should preferably encompass all visible colors. It turned out that in order to do so, the primaries had to be defined OUTSIDE the visible colors. So they defined imaginary primaries which tightly encompass the visible colors but are not themselves part of the visible colors. These primaries were then called XYZ.

There is no such thing as a reference CMYK because subtractive spaces as we know them vary wildly in behavior and certainly do not act linearly at all. So there is no useful relation between CMYK spaces, and writing something like C+M+Y=black is not applicable in all practicality, nor does it make sense from a modeling point of view.

For relatively stable CMYK processes such as press printing it certainly is possible to define an output standard, thereby allowing input with predictable output. I doubt that there is such a thing as a "one-size-fits-all" conversion between that input and output, but current software goes a long way in providing the controls required by most prepress needs. With current technology advances, this may well become interactive control in the near future, allowing dedicated, per separation, predictable conversions.
Logged

Regards,
Oscar Rysdyk
theimagingfactory
Henry Goh
Sr. Member
****
Offline Offline

Posts: 574


« Reply #13 on: May 17, 2006, 03:49:26 AM »
ReplyReply

I have followed the "original" thread as well as read this one.  I'm just a photographer who needs to send digital files to my local printer.  Most printers here (small shops) know nothing about colour profiling, calibration etc and so they will only work with proofs from the colour separation house and try to get as close to these proofs.  Now that separation film is being eliminated and direct-to-plate is in, the same printer has bigger problems.  he continues to depend on contract proofs to guide him.

As a photographer can i ask you expersts here if I'm on the right track:

I work in PS in RGB, mostly Adobe RGB (1998) or ProPhoto RGB.
When I'm done, I convert to something like a "standard" sheet feed V2 profile and save the CYMK TIFF.  I then send this file to the plate maker and he outputs the plate with a Matchprint type prrof for the printer.  Considering what I have described earlier about the type of printer I'm working with, would you say this is a reasonable workflow?

Also, when I'm in PS Adobe RGB, should I softproof using the Sheet Feed V2 profile and then adjust for tonal and color saturation etc before i convert?

Thank you.

Henry
Logged
opgr
Sr. Member
****
Offline Offline

Posts: 1125


WWW
« Reply #14 on: May 17, 2006, 04:57:38 AM »
ReplyReply

Quote
Considering what I have described earlier about the type of printer I'm working with, would you say this is a reasonable workflow?
Considering the circumstances and presuming the platemaker doesn't do colormanagement either, then what you describe is a reasonable workflow, But only if you can make a reasonable translation between your image preparation and the matchprint from the platemaker, and only if you prepare for relatively default print methods.

It would however be strongly advisable to search for a colormanaged production chain and see how that compares to your current experiences. There are some major advantages:

1. You can transmit the responsibility of separation to the next link in the chain. A prepress house may be able to optimize the separations for different print-houses or print-methods. If something goes foul, it is not your responsibility. responsibility and expertise should go hand-in-hand as in: if it is not your expertise, then it is best if it is also not your responsibility.

2. If any of the variables chance, it does not impact your workflow (or responsibility).

In either case, working with your current setup or in a colormanaged chain, it is useful to add something like an Epson proof to your own workflow.


Quote
Also, when I'm in PS Adobe RGB, should I softproof using the Sheet Feed V2 profile and then adjust for tonal and color saturation etc before i convert?
It is ALWAYS advisable to softproof and optimize for the final output. Preferably before conversion, so that you don't have to redo the corrections for additional conversions.
Logged

Regards,
Oscar Rysdyk
theimagingfactory
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad