Trouble Following CMYK Separation Procedure

[Jeff F]

On page 139 of Richard's book, while attempting to create the Luminosity Mask, I'm not getting anywhere near the result I get when choosing the "CMYK Black" tool from the Hidden Power Tools menu. When I accomplish steps 4 and 5 in the book, I get a 50% gray mess as the "Black" layer. It's impossible to tell that there's a man in the photo. However, if I move the Saturation layer above the Luminosity layer then merge these two layers down to the "Black" layer, I get a result that looks a lot more like what I get when just using the tools menu. Perhaps I'm misreading how to do this. Certainly I'm having difficulty understanding how the luminosity mask and the saturation mask work together to create the "Black" plate. Anyone who can shed light on this issue please feel free to comment.

Thanks,
Jeff

[Richard_Lynch]

You need the luminosity to focus on the darker part of the image, and the saturation to focus on the grays and blacks rather than colors. If you were to mix black into colors they would tend to get muddy. The Black plate is technically only supposed to affect the very darkest of colors.

Luminosity masking alone won't do it.

That help?

[Jeff F]

Thanks again for helping me out here, Richard. I believe that I understand now why we need the luminosity and saturation masks. However, in my original post I was more concerned about whether or not there is a mistake in the steps of the book or if I'm misreading it somehow. It says to move the saturation layer above the Black layer then merge the two. This leaves the luminosity layer out of the mix -- it's still on top of the stack but uninvolved in creating the Black layer it would seem. But if I move the saturation layer to the top of the stack and merge down to the Black layer, the luminosity layer is sandwiched in between and is part of the solution. It also looks more like what you get if you just use your tools from the CD. Is this a mistake in the book? I'm just trying to get the best understanding that I can from this very difficult subject matter.

And since you brought up the notion of muddying the colors with black added to the CMY separation, even after using the Hidden Power tools to accomplish the CMYK separation on Gorskii's photo, the result muddies the colors a little as compared to the original. Is this simply a trade-off that one has to accept for adding black ink to the mix? Is it possible to get an exact separation that looks identical to the original while using a black layer? Or is this just a problem with showing a CMYK image on an RGB device (monitor). Sorry, but I've been unable to try out the actual printing of this yet. I am looking forward to it, though. I played with it onscreen for quite a while last night to try and figure it all out, but I was unable to get colors as bright as the original without deleting the black layer altogether.

Any thoughts on these subjects would be greatly appreciated,
Jeff

[Richard_Lynch]

RGB is not the same as CMYK, as it says in the book. RGB and CMY will look identical on screen, but the print will notably fade. Black has to be there to emphasize the tonal dynamic you won't get from CMY inks alone.

About the steps...Step 4 does say to move the saturation layer above the black layer. This will lighten the result of the black layer -- effectively masking the saturated colors for being tainted with much black.

I'm pretty sure those steps are correct -- the tools were built to mimic them. If you have the HPA2 action player, you can slow down the playback speed and open the layers palette to watch the steps occur.

That help?

[Jeff F]

Ok, I think I found the problem. I slowed the action down, and it appears that the Luminosity layer it grouped to the Black layer (actually 50% gray), changed to a luminosity blend mode, then merged with the Black layer before messing with the saturation layer. I don't have the book in front of me, but I believe this step is missing from the book. If not, I apologize for wasting your time, but I read through that section a couple of times and didn't see any mention about the luminosity layer other than a levels correction. Boy, if I missed that, I'm going to have to talk the doctor into increasing my medication! But now that I've seen the action slowed down, maybe I can start to get my arms around this monster.

Thanks for your patience,
Jeff

[Susan S.]

I've finally bit the bullet and attempted to work through this section again, and I think Jeff's right. if I follow the instructions in the book the luminosity layer gets left sitting doing nothing - step 3 on p139 finishes the luminosity mask, but it isn't mentioned again, and isn't grouped with anything to produce a mask. The merging in step 5 only talks about the saturation and black layer being merged. (as I recall this is where I got stuck last time, and thought I must be doing something wrong as I didn't really understand what it was trying to do - I'm a little closer now!)

Actually there are surprisingly few errors in the book for so many detailed steps. And I think this is the first one that has totally stumped me. I'd hate to have to proof-read/test something like this....(my husband's law texts are bad enough!)
Susan S.

Edit: to add - I've just run the actions for the CMYK Black tool - this does appear to have other differences to the instruction in the text as a curves dialogues appears (I can't find any reference to curves in the text version...I guess I need to step by step thru' it to find out but I'm suffering a caffeine deficiency this morning! - my guuess is it is in place of one of the levels adjustment layers)

[Jeff F]

Hi Susan,

I'm glad it wasn't just me getting lost at that point. I, too, noticed the addition of the curves layers after slowing the action down. I believe they are in there to make the tools more complete for those with expertise to be able to fine tune their results at each step of the separation. My guess is that Richard left them out of his manual steps so as not to add difficulty to an already complicated procedure. He does, however, mention their availability throughout the text for the separation for making precise changes to the different color layers. I must confess that I have struggled with the black layer part of this separation, and even now am only on the most rural outskirts of understanding. But that will probably have to suffice for now at my current experience level.

Good luck with those law texts,
Jeff

[Susan S.]

Jeff - I'm now at the stage of understanding what the CMYK separation process is doing and the role of the different masks- I think. the next step is getting on top of the steps that are used to create the masks and understand that fully.

I really ought to go back to reformatting chapters that came back from his co-author with all the italics removed and in the wrong font and with all the style sheets scrambled (why do law books need that many footnotes?!!)....but right now CMYK is more interesting!

Susan S.

[Jeff F]

Susan - I'm glad you're getting a handle on the CMYK process -- that's one more person who can answer my questions. As for me, I know I need to devote more of my energy toward gaining experience in more basic correction processes. But I'm just too inquisitive, and it will bother me until I have a firm grasp. As long as I keep making forward progress, I'll be alright. And thanks to Richard's book and all of the help I've received here, I'm headed in the right direction. (Note to self: Crawl first, then learn to walk).

Jeff

[Richard_Lynch]

Finally had a chance to check this. You should merge the Luminosity and Black channels between step 3 and 4 -- there is indeed a step missing. If you try it with the tool, you can follow along in the steps and it becomes pretty clear what is going on. The Luminosity is used to generate tone for the black -- it is then corrected so that the black only is introduced at 50% or darker tones. It is merged with the black layer to do 2 things: change the mode of the layer and change the name to Black.

I'll have to get that in the errata. Thanks!

[Susan S.]

Thaks for that Richard. (Did you get my email? - my mail server is playing up again and some stuff appears to be disppearing into an electronoc black hole!)

Susan S

[Stephen M]

Although I wrote these articles for the full version and not PSE - they are a small attempt to help explain some of the theory behind working in CMYK in Photoshop. I hope that they are of some help, please ask away if any clarification is needed on the theory, but I can't help on the actual PSE side of things.

http://members.ozemail.com.au/~binar...T_CMYK101.html

http://members.ozemail.com.au/~binar...T_CMYK102.html

http://members.ozemail.com.au/~binar...T_CMYK103.html


Hope this helps,

Stephen Marsh.

[Richard_Lynch]

Stephen,

The problem with those articles/descriptions is that they apply to terms having to do with settings in Photoshop. None of them are available in Elements. The idea of CMYK separation and generation of color plates has to be handled entirely manually from an RGB file. Your articles and replies deal with screens and settings that don't apply at all to the Elements user creating a separation -- the terms describe things that come into play, certainly, but there is no real way to look at them as flat settings or simple considerations. CMYK functionality as you understand it and as it exists in Photoshop are not part of Elements.

The solution in Elements has more to do with the *art* of color separation as it was approached BEFORE there was digital pre-press. There isn't a CMYK mode in Elements.

One quibble with the articles, unless I misunderstood, there is a third way to separate: manual digital separation saved to a DCS file. That's what we do in Elements -- or at least it is the only way I have currently found to accomplish this, and the way it works in my book. I'd think you'd find it interesting...Quite a bit of control over the result, really, in a way you can't imitate with Photoshop's controls -- and limitations -- without using similar technique.

[Stephen M]

Richard - I thought I made it clear from my disclaimer that the articles were written for the full version and that PSE users would need to figure out the differences as I could not help there. But perhaps not, I guess I should have been more clear. <g>

My site and the articles in question are written from a certain biased viewpoint - being prepress, which is something that I spend a lot of time doing. Many Photoshop users do not understand CMYK and that there are many 'flavours' of CMYK that one may wish to create - be it different separation types (UCR or GCR) or different targeted separations, say for coated flatsheet or uncoated web newsprint or the other variables such as reproducing the same colour on two similar conditions that use different ink - just the day to day stuff that is needed for the serious CMYK user.

Some do not understand the relation of CMY and K to neturals and colours, which I attempt to go into with UCR and GCR. There are many mixtures of CMY and K that one may wish to produce. Sometimes no or little black is wanted, othertimes more is required. Sometimes you want black in coloured areas other times only in the deepest shadows. It all depends. Again I attempt to go into this, and it matters not how the separation is created, some things are fundamental and the info is applicable to all, it is up to the user to make their tools do what is required. For example, one part mentions that more K in the sep is good for a four colour greyscale - so all the PSEHP user needs to do is figure out how to use your tools to create a separation with a heavy black plate and lighter MY and C plates (the C being a bit heavier than the MY to form a true neutral). Basic separation theory is the same no matter the tool - one just has to figure out how to perform the required edit with the tool at hand. As previously mentioned, many full version users think that CMYK is CMYK and that their are not different variables which one may use to create said CMYK. This is not helped by many in the print industry simply saying "just give me CMYK" instead of saying "give me CMYK that is suitable for SWOP TR001 type reproduction". <g>

It can be an eye opener for many that one can target different stock or ink or dot gain or other variables. Many think there is a magic 'one size fits all' CMYK space that is suitable for all uses. So some time is spent with the legacy separation controls which have fallen from favour with ICC profiles but which are still very useful. Then ICC profiles are explored, which have their own set of rules for working with them.

The reason for posting was a small attempt to bring the wider use of CMYK into focus. It is helpful to understand why one would like a certian type of CMYK over another and why one type of press may have the same image separated in many different ways. It was more about the bigger picture than the steps which may not apply to PSE users.

You are correct, due to PSE not having a CMYK mode users of PSE have a certain advantage over their full version cousins - in that they are forced to learn something for which the regular version users take for granted as being performed by a separation table or ICC profile. When it comes to spot colour work, many folk are lost as they have lost the art of creating separations - due to the fact that Photoshop handles it all for them when they do CMYK.

As the articles are written from a prepress biased viewpoint, where consistent and accurate reproduction for many output conditions is required, performing the separation as a PSE user is forced to is not even an option for most work for most users.

Full version prepress users have often manually created seps for grey, spot or CMYK - but even a die hard manual separator can't compete with the workhorse of modern separation sofware and as colour is no longer an art but a commodity there is not much profit to be made in performing manual separations. Yes it is an option, but not a serious one for most prepress settings. A PSE user using CMYK for their own sake who has the time to devote to their work can justify this. A stressed prepress operator having to separate a few hundred images in a short deadline for a crappy rate is not going to trot out PSE and perform manual seps. They are going to run a batch action or AppleScript and if lucky each image may have the K plate massaged a bit or some post separation selective colour moves applied or something quick and pleasant but nothing time consuming.

Stephen Marsh.

[Stephen M]

In addition to this thread:

There is a way for PSE users (or anyone) to create a CMYK file using ICC profiles - but they do not need PS Elements or even the full version of Photoshop to do it!

On the Mac, there are ColorSync AppleScripts for performing _many_ ICC related tasks, including converting between profiles (such as RGB to CMYK, or more specifically say Adobe RGB to SWOP TR001 CMYK).

On the PC there is some free command line tools for this - but I have never got them to work, as I am a old Mac user and have no experience with command line operations. If one can understand the syntax that these apps use, then one can do ICC profile based transforms to TIFF files and a few other things.

Whether or not PSE can open or use the CMYK file you create with these apps or others is something else again, but as I mention there are two built in ways to separate in Photoshop as well as manually, and with Richards tools there is a manual way to do it in PSE.

Stephen Marsh.

[Richard_Lynch]

Stephen,

One should walk, or at least crawl -- or scootch -- before they run. The idea of Elements use for CMYK is that you can do it, though the manufacturer says it is impossible. Meanwhile, you learn a lot about how images perform and learn about opportunities to control output. Throwing in 100s of variables into the mix for the sake of theoretical and actual possibility is not the point or the desire -- and neither is using elements for high-volume production CMYK (pre-press professionals will probably not be seeking out elements to do CMYK work). One of my few stipulations for those needing Photoshop is that they will have to want one of these things; to work in 16-bit, to do high-volume production and CMYK, or to record actions. The suggestion is: if theere is need for a specialized professional tool for volume work, one will probably be using Photoshop. It is the justification for the cost difference.

While I think your expertise could do some good here, dismissing the tools that the Elements user has to work with may find the user confounded, overwhelmed, and confused. It is hard enough, I think, to digest the simple idea of manual separation (why CMY does NOT work for RGB). If you'd like to discuss methods of control for factors as a process in separation and as part of using the available tools (not merely saying "Adjust for dot gain.") this would prove more valuable. You do not have dot gain settings, you do not have transfer functions, you cannot embed profiles (in these images), you do not have GCR or UCR settings. The black is generated from a combination of saturation and luminosity masking, and while this can deliver very fine control of all aspects of the final result, tossing in these terms is really meaningless in the context of the interface. How they can be calculated in terms of using image information with curves, masking and calculations (or other functions) would be germane.

[Stephen M]

Thanks Richard, I understand the intention and that you have come up with a nice way to work around a limitation of Elements that can be helpful to some users and situations.

Yes, PSE will not be the first thing a prepress user will look for - but having a stripped down version of Photoshop is something prepress folk would like, with features targeted to them and throwing away all the other unwanted deadwood, so there is nothing wrong with a less featured product in my eyes. It is not the amount of features, but what they actually are and how well they are implemented that concerns me. I could care less how well audio annotation works in Photoshop, I have no use for it. <g> But the USM filter interface and results are a big concern.

I mentioned PSE and manual seps because you said that is the only way to create seps in PSE and that I had neglected it as an option in my articles aimed at full version CMYK users - my point was that one could use PSE or the full version to manually create seps too, but most CMYK users do not have the time or budget to manually perform tasks that have been automated to high degrees of accuracy and repeatability. Yes it is a valid option and is still performed by some folk today who have Photoshop, but not for the majority of users or work in settings where press output is concerned. The separation methods listed in my article were in context to that situation but not to elements, which I originally noted before the links were supplied.

I do not think I have dismissed the tools or PSE out of hand, but for some specific work it would not be ideal, as you have noted. Manual seps may not be what a full version CMYK user would ideally choose - but that is not to say that there is no merit to them or that they may not be very useful to some folk. Being forced to manually separate CMYK would help many users - including me. But that is not going to happen at my day job, but it would be a nice exercise for play time in Photoshop.

If a PSE user is going to make their own CMYK seps, and their intended use is commercial printing - then they are going to be confronted with terms that do not apply to PSE, unless the printer is doing the conversions from RGB for them (which would be a good idea). Photoshop did not invent most of the terms or needs of press and prepress, it just responds to them (many of the terms predate digital separation).

So, with luck a PSE user that knows that there are such terms as GCR, UCR, grey component, total ink limits, dot gain and other jargon may be able to look at what factors are needed for successful CMYK generation and may then further research what these functions do and if they can/should be emulated with the tools that PSE offers.

The reason for posting the article links was to add to the general topic of CMYK, more so than the tools used to get there. In my earlier post I did mention that I was more than happy to expand on the theory behind CMYK so that a PSE user could try to translate that into their tools. The offer still stands, but without direction I can only use the current three articles that I previously pointed to for explaining some of the things that affect CMYK generation/use.


Stephen Marsh.

[Jeff F]

Thanks, Stephen, for posting the links to your articles on CMYK. They're very well written and informative, and I think I absorbed quite a bit from them. It would appear that CMYK separations provide much more control than I am ready for. It must be akin to piloting an airliner. Most business travelers would probably like to have more control over their itinerary, but I doubt that flying in the captain's seat is what they would have in mind. I'm afraid that I will be relegated to a passenger seat for the time being while my print drivers do the flying.

However, that's not to say that I'm not fascinated with the subject. I'd very much like to hear from anyone using Richard's method of CMYK separations for printing on a home inkjet printer to understand the control they're looking for and how well they have been able to achieve it. Have you noticed any distinct advantages or disadvantages?

Thanks,
Jeff

[Richard_Lynch]

Jeff,

As I was trying to say -- and my worries about Stephen's posts -- there is no reason to get scared off...

While there can be a lot of complications, if you follow what's in the book (or use the tools!) you can easily try it without adjustment from what I've preset. If the result is not to your liking, there are ways and means built into the process as I've set it up to allow you to adjust the result.

[Jeff F]

Hi Richard,

In all fairness to Stephen, I don't believe that anything I've read from him has scared me off of CMYK. I had reservations about the utility to me of that particular separation before Stephen chimed in. That's why in my last post I ellicited some anecdotal experiences from other users on how they're using your procedures to their advantage when printing from a typical inkjet printer. Unfortunately for me, I have still not reached the printing chapter of your book (I'm now on chapter 7), so I have not had a chance to play with it yet. Perhaps I should hold any other questions on this subject until I have reached that chapter and have played with the prints. Many of my questions may be answered there.

Thanks to you, Stephen, Susan, and everyone who take the time to share experiences and opinions with amateurs like me.

Jeff

[2003blink]

This problem has not made the errata list as of yet. I just spent 2 days trying the steps over and over and over. Throwing the book, cussing Richard (not really Richard ) and taking more aspirin. Of course I tried the easy route first, you know, read the eratta...

[Richard_Lynch]

blink,

And what problem might that be? I am not clear. You can do CMYK separations in Elements. There is one error I know of, but i am not sure which you want addressed in the errata.

Stephen,

The idea would be to use CMYK in Elements in a casual and non-production way. It would also be a great experimentation ground for one to find ou about and work with the nuance of CMYK adjustments WITHOUT the benefit of Phooshop automation so there is an understanding of what is going on and what needs to be achieved.

My opinion is that users need to learn this before being really effective with separations. In your case you already know the theory (via having to create separations and/or fighting with them). I refuse to spare the user the glory of understanding in favor of pushing a button.

Richard

[2003blink]

Sorry Richard, I had just finished reading the whole thread, and still had this fresh in my mind. I should have quoted you, my mistake. The problem is the missing step (s). ALSO please note that on p140 "Remove Color Under Black"; step 1. Copy the Black layer and Invert it.... Probably should say Duplicate the black layer and change the mode to difference.

I figured out the copy (lol), but used cmnd+I to invert. I don't think invert was the correct step, and am pretty sure mode to difference is the correct step.

I left my book sitting for now, as I got very frustrated with this at this point.

>>>Richard_Lynch
Moderator

Registered: Oct 2001
Location: Western NY
Posts: 244


Finally had a chance to check this. You should merge the Luminosity and Black channels between step 3 and 4 -- there is indeed a step missing. If you try it with the tool, you can follow along in the steps and it becomes pretty clear what is going on. The Luminosity is used to generate tone for the black -- it is then corrected so that the black only is introduced at 50% or darker tones. It is merged with the black layer to do 2 things: change the mode of the layer and change the name to Black.

I'll have to get that in the errata. Thanks!

__________________
Richard Lynch
http://hiddenelements.com/<<<