Skip navigation
This discussion is archived

This Question is Answered (go to answer)

6,505 Views 32 Replies Last post: Jul 3, 2011 3:56 PM by walton RSS Go to original post
Level 5 9,257 posts since
Aug 22, 2008
Currently Being Moderated
16. Jun 30, 2011 12:22 PM in response to: firebuffgal
Re: Have you printed Color Photographs in your book?

My understanding is that you can't really calibrate a digital press to a specific profile -- you can only calibrate it back into a "known state" that's unique to itself, then characterize that (an ICC profile) so it can emulate a given industry standard.

 

A custom printer running a digital press will have a reliable profile for that device, they'll give it to you for soft-proofing, and show you how to prepare files for a color managed workflow that makes complete use of their capabilities.  In contrast, LightningSource takes the approach of discouraging color management, and allows only DeviceCMYK color.  However, LightningSource also states that HP Indigos are the backbone of their on-demand color printing operations, and they tell you what kind of paper they use -- so you can find, install, and apply a relatively appropriate CMYK output profile (essentially letting you manage the color on the front-end) in order to get predictable output.

 

CreateSpace doesn't really tell you anything about anything -- so each of us who've done color projects have found (or are still finding) the techniques that work best for us.  Unfortunately, there are inconsistencies in what we've seen firsthand.  (e.g., I've seen CS color books that were printed on an Indigo, but I've seen some that weren't.  What I've seen in terms of screen frequency, halftone shapes/patterns, and page imposition don't match what walton's contacts at HP and CS insist are being done.)  Further, we aren't necessarily working with the same types of source images, may not be dealing with the same challenges or problems, and may not have the same objectives for our projects.

 

For example, collaborative work is one of my big challenges, where four different people opened and manipulated an image before I ever saw it (one probably using Paint or some other software that had no color management capability) and they totally hosed the source color tag.  The text says the little girls are eating popcorn and drinking lemonade, but they look like Lady Gaga with guacamole and margaritas in the image I received.  I'm not trying to preserve those out-of-gamut colors -- I'm just trying to get it into some colorspace that is reasonably accurate, that the collaborators can soft-proof reliably in the PDF, and that CS, Amazon, and the partner printers can produce such that there are no surprises.  Naturally, my "best practices" for that project might not be particularly useful for an art quality book where you had end-to-end control over the native images.

Level 5 14,225 posts since
Sep 5, 2009
Currently Being Moderated
17. Jun 30, 2011 1:15 PM in response to: Lighthouse24
Re: Have you printed Color Photographs in your book?

I am baffled: my conclusion is that if art is prepared in RGB and submitted in RGB, CS will print it "dead on" in terms of perceptual color fidelity, but if the art is converted to CMYK the colors will be off, unless you spend the time color correcting, and you probably won't be any better than the RGB version, but you will have wasted hours and hours. 

 

If CS can print 5 identical color references, each tagged differently (Apple RGB, Adobe RGB, PhotoPro RGB . . .) and they are the same in print, how does that alter my conclusion? How would you expect them to print?

 

I would expect them to print the same. So what?

 

I have said and will say again, CS strips away (some or all--I don't know) embedded color, intent, etc. information.  It must.  Most of us in the US using Photoshop keep many or all of the defaults for color management (North American General Purpose 2, sRGB (etc.), Gray-20% dot gain . . . ) but:

  • CS changes the dot gain (Indigo uses 9-12% . . . but I don't know what CS uses because their paper would be specific)
  • CS changes the GCR (Indigo said)
  • CS changes the UCR (Indigo said)
  • CS would change some of the things the ICC standards talk about: Black Point Compensation, colorimetirc rendering and rerendering intents, white paper reflectiance, etc.

What does CS keep or strip away? What does CS keep? What criteria control what?

 

My goal has been to develop a "profile" (simple, visual, non-technical) of CS's printing capabilities. We won't get their profiles. And I am trying to keep this simple.

 

[The following is all in italics because the system made it in italics, which cannot be removed--I removed the html code, but it keeps being put back!]


I have been trying to reconcile your suggestions here with why in Color Sample Reference, the RGB values were given for the SVG Named colors (CSS Color Module 3) , and were also shown in CMYK values (I have seen this list elsewhere, but I have no idea how it was derived, what intents were use, etc.: because the RGB (CSS Color Module 3) values  match in CMYK values (Color SVG in CMYK).  I've have read over  http://www.color.org/ICC1v42_2006-05.pdf for a simple explanation for the actual CMYK values shown as equivalent to the RGB values, but nothing is simple.)

If you can explain this and why you converted the RGB examples so that they no longer are what they say they are, in simple terms, I'd appreciate it.

All color samples were created using the CorelDRAW Graphics Suite. . . . Those RGB samples were then converted to CMYK percent age values using a conversion tool provided by Peter Forret . . . . All three samples should look identical . . .  [page 3]

[presumably Peter Forret refers to http://web.forret.com/tools/ , or similar . . . I looked for an explanation regarding converting these web colors from RGB to CMYK but couldn't find it, which does not mean he does not give one]

This is my problem: my first concern was determining CS's actual ink colors. The Photoshop Colors (below, top) show both the CMYK values. . .
(which both charts show as process color values--for example, my Process Cyan C100 M00 Y00 K00 matches the stated values in Color
Sample Reference for Aqua or Cyan, C100 M00 Y00 K00, etc.)
. . . and the RGB values based on the Color Sample Reference charts decimal color codes. As predicted, CS printed the SVG & CMYK values the same in Color Sample Reference (below, bottom) because the RGB values were changed, see quote above. But, if most people work in RGB and that is what CS prints from, why change to CMYK when the values will be different?

 

 

As can be seen, in Photoshop the CMYK colors are very different from the SVG colors.  In CS Digital when CS printed RGB versus CMYK colors, similar to those in the example, the colors printed as distinct color, as in the example above. Kind of like, what you see is what you get in RGB, so why convert it? Why show RGB colors that no one can make from the RBG decimal code without a secret decoder ring?

 

I trimmed out the Aqua or Cyan example (Color Sample Reference, page 5)  and taped it to my CS Cyan, (C100 M00 Y00 K00, from page 35 of CS Digital). I scanned these at 1200dpi simultaneously. In Photoshop, I then took a patch and blurred it to make getting an accurate reading easier:

 

 

Photoshop color readings:

  • (left) Color Sample Reference cyan C76 M12 Y20 K00;
  • (right) CS Digital cyan C74 M18 Y0 K0. 

Are these close enough to be explained by a variations in ink color?  Or is the difference an artifact of converting to CMYK before printing (left)?

CS Digital was printed in Charleston, SC, Color Sample Reference was printed in Lexington, KY.

 

Under a loupe, but not a microscope, the Color Sample Reference cyan (left)  appears to have more of a dot pattern than the CS Digital sample (right). At 800% there was a regularly recurring yellow pixel in the sample on the left.  Guess I'll have to buy an microscope.

 

I read most of Gary Ballard's site, http://www.gballard.net/psd.html. Interesting, and I have no problem with any of it. But CS isn't playing by the rules.

 

To anyone reading this . . . after the first sentence, it is all hot air!  CS does a great job of color fidelity. Stay in RGB. Don't read this junk.

 

Walton

Mechanics & Punctuation, free, 20 page guide to everything punctuation  Build Your Book,  free 98 page guide to designing your book  CS Digital  understand CS digital possibilities GIMP, free, tutorials, GIMP, GIMP Help, excerpts from GIMP Supremacy Supremacy  Contact for graphics, design, and typesetting help  Disclaimer: all statements of apparent fact in this post are empirical inferences based on observational data. These are idiosyncratic in nature and have not necessarily been subject to verification.

Level 5 14,225 posts since
Sep 5, 2009
Currently Being Moderated
18. Jun 30, 2011 2:35 PM in response to: firebuffgal
Re: Have you printed Color Photographs in your book?

There are industry standards, but how many printers (non-digital and digital) follow them to the letter?  Is that even possible iin all cases?

 

Indigo has it's own proprietary inks and profiles. Profiles are great.  Link the press, ink, paper as the output to the input (digital file) . . . which in turn is linked to the original user's monitor, printer (ink and paper), and Photoshop.  What I see at one end is what I get at the other. Or is it?

 

An analogous aside: I watched the sensitometry people kick themselves silly matching developer, paper, lens, filters, camera, film into a solid brick of control.  I asked a lot of them . . . at the A level . . . about shutter speed. Did any of the them test their lenses, and if so what did they find?  I tested dozens of lenses at dozens of aperatures . . . plus or minus as much as 200%.  Predictable? No.  Some lenses were perhaps conisistantly in the -50% to +75%, or -25% to +50% . . . change the aperature, different range. You could test each batch of film, paper etc. I don't know anyone who tested the enlarger  light with a colormeter or other measuring device.  The point is, it only could go so far. As far as I could tell, it ultimately came down to the printer. Sensitrometry was not what Zone VI promised.

 

If as several experts have suggested to me, CS, with 15 years of experience, has probably developed profiles for the work it gets most often. That's what it uses.  Probably sRGB (I usually just say RGB).  Strips what it doesn't want, replaces what it does. Most members are happy, and most member wouldn't know  or care about any of this.

 

If you want a known profile, or if you want a digital printing company to run proofs for approval, with you (the client) there, tweaking the curves until you are happy, that's a piece of cake for the Indigo, lots of companies do it, but it's not what CS does.

 

If I've messed up anyone's work with some of my comments, let me restate what I've observed: If you work in RGB mode and if you upload RGB, you'll get CS's best color printing.  "Work in" means also doing the things that make images work: levels, curves, selective color treatment, etc. I assume that's a given.

 

Walton

Level 5 14,225 posts since
Sep 5, 2009
Currently Being Moderated
19. Jun 30, 2011 3:22 PM in response to: walton
Re: Have you printed Color Photographs in your book?

Look like I messed up a graphic. The SVG yellow in the graphic above is R255 G255 B00.

 

Walton

Level 5 9,257 posts since
Aug 22, 2008
Currently Being Moderated
20. Jun 30, 2011 3:40 PM in response to: walton
Re: Have you printed Color Photographs in your book?

walton wrote:

If CS can print 5 identical color references, each tagged differently (Apple RGB, Adobe RGB, PhotoPro RGB . . .) and they are the same in print, how does that alter my conclusion? How would you expect them to print?  I would expect them to print the same. So what?

 

Ah, that's why I asked you to try it.  When I did it, they didn't print the same.  AdobeRGB was the only one close to correct.  Does that mean CS honors only one color profile and ignores the rest? Does it mean they assign AdobeRGB to any RGB content in the file?  Does it depend on something I don't know about?  Would you get a different result than I did?  I don't know -- and I didn't need to know (the fact that a tagged color profile wasn't honored, even if it was only once, revealed what I was trying to determine).  Since you asked for things to test, though, I thought I'd suggest that.

 

I'm waiting for a flight at the moment, and having a little trouble concentrating on and understanding what you're asking about the Color Sample Reference.  I'll have to come back to that later when I'm in a quieter place.

Level 5 14,225 posts since
Sep 5, 2009
Currently Being Moderated
21. Jun 30, 2011 4:14 PM in response to: Lighthouse24
Re: Have you printed Color Photographs in your book?

Then I'll put in 2 pages, tagged differently and see what happens.

 

But I hope'll come back to my question.  It's been baffling me for weeks.

 

Walton

Level 5 9,257 posts since
Aug 22, 2008
Currently Being Moderated
22. Jun 30, 2011 10:22 PM in response to: walton
Re: Have you printed Color Photographs in your book?

walton wrote:

If you can explain this and why you converted the RGB examples so that they no longer are what they say they are, in simple terms, I'd appreciate it.
But, if most people work in RGB and that is what CS prints from, why change to CMYK when the values will be different?

 

Kind of like, what you see is what you get in RGB, so why convert it? Why show RGB colors that no one can make from the RBG decimal code without a secret decoder ring?

My summary answer would be that you're trying to make my booklet into something other than what it is.

 

Go back two and a half years (when I started it) -- the posts in this community that discussed color were more along the lines of "My cover is red, but it looks pinkish-purple on my first proof and sort of brownish on my second proof, but I didn't change the cover file.  Why did that happen?"  A lot of the responses were along the lines of "CS/digital presses can't print color accurately or consistently."

 

I knew that at least Lightning Source and two other short-run book printers I'd worked with could produce accurate and consistent output on digital presses from PDF/X-1a compliant files -- at least for covers (at that time, I hadn't done any books with color interiors, and we hadn't acquired a design firm yet).  So my objective back then was to determine if this really was a limitation of CS's.  My question was, if I prepare a file the way I would for any other printer, will the color be all over the place as suggested (with variations from book to book).

 

To test that, I wanted to start with a named color (e.g., "red") that had an accepted/standard RGB value (so anyone could look it up or match it on a color chart), then produce a CMYK swatch that matched when I soft-proofed it, when I printed it locally, or when I compared it to a 4-color Process Guide, and finally see if the red that CS printed in the booklet was (1) the same as a commercial Process Guide sample, and (2) the same from book-to-book.  So to answer your question in simple terms, I converted to CMYK because I was preparing a PDF/X-1a file.  I was preparing a PDF/X-1a file because every single printer I'd worked with prior to CS required it for color covers or books, and the results had always been consistent.

 

There is nothing in that booklet or its description that presents it as a tutorial for how anyone ought to handle color on his or her project.  Its purpose was to show that you can start out with red (or any of the other 138 named colors I used) in a source file and end up with almost precisely the same color in a CS printed book.  So if someone's red cover is pinkish-purple in the first proof and sort of brownish in the second and they didn't change anything, there's a pretty good chance that neither CS nor digital press technology is to blame.  I didn't say PDF/X-1a was the one and only path to that outcome -- only that it was the path I took.  I don't promote the booklet here in a signature line or in my profile.  I only offer it to members (with a discount code) when they ask for samples of color books and/or express a concern about the quality of CS's printing and the ability of a digital press to produce acceptable colors -- because the booklet visually addresses that.  It's not a how-to guide.

 

Today you write, "CS does a great job of color fidelity" -- but that wasn't the prevailing consensus here when I started the project.  I hope that puts things in perspective and answers your questions.

Level 3 573 posts since
Sep 24, 2009
Currently Being Moderated
23. Jul 1, 2011 6:52 AM in response to: Lighthouse24
Re: Have you printed Color Photographs in your book?

Now I'm curious.

 

In the case of supposed-to-be-red-but-is-sometimes-pink-and-sometimes-brown when there have been no changes in the source file, what does cause this obvious color shift other than printing techology issues or excessive variation in the printing process?

 

("Sometimes pink" and "sometimes brown" do not seem to be on my list of named colors, but "red" is.)

Level 5 14,225 posts since
Sep 5, 2009
Currently Being Moderated
24. Jul 1, 2011 7:15 AM in response to: Lighthouse24
Re: Have you printed Color Photographs in your book?

Thank you.  I did understand your answer: you used SVG colors, which were set up as a common standard for Web graphics.  I was thrown off by "This allows CreateSpace Community self-publishers to see precisely . . . . By comparing the colors in this booklet to (a) the same colors as rendered on your computer . . . "

 

I was not trying to turn your book into something it was not. I understood its purpose for you, but not for the "CreateSpace Community." I simply don't know enough to know how to use it as a resource.

 

In another post, I said that CS did not use standard/traditional process colors, and you said or implied CS did (somewhere, literally or in the context of the post, based on color values if not from  your book, at least similar to them).  The SVG colors (in particular: Cyan, Magenta, and Yellow) are not PMS (or similar standards) printers' colors, as my color examples above suggest. And if there is a second or third color in the Aqua or Cyan, page 5, Fuchsia or Magenta, page 17, and Yellow, page 21, (all given with pure C, or M, or Y values), as I suspect . . . this says something curious about both the art, the press, the purity of the examples.

 

My experience regarding color printing 2002-2011 from BookSurge and through CS, has been that aside from all other problems, it's color fidelity and consistency have been stunning. That people compained about it, a consensus that I was opposing at the time, is why I set out to test CS's digital capabilities . . . and why I have come to the conclusions I have.

 

On a separate note, I have inserted two test pages concerning tagged color profiles. I have use my two color references so that they tie into the rest of my tests. They include 10 different flesh tones, a color ramp, and the ColorChecker with the original RGB (sRGB) PDF art available on line for a true art-to-print comparison. I've embedded CIE, Apple, Adobe, ColorMatch, PhotoPro, sRGB RGB profiles, with an identical panel made up of screenshots of the same art, showing the different color values but all in one sRGB file, so there is a fair comparison before and after CS prints them.

 

I have been thinking about what I said yesterday, and I think I was wrong: I have been saying CS is set up to print RGB files better, more accurately in terms of perceptual color values, than from CMYK files.  Demonstrably true. But what is the most common RGB profile for scanners, cameras, monitors (I suppose I have to add telephones to the list!) . . . arguably sRGB. And my tests and experiences have been made using sRGB. It will be curious to see what CS actually will do with different RGB profiles, but they could honor the profiles or alter them, either way this still conforms to my premise, but it is getting into realms that 99.9% of CS members could care less about, and truly don't need to care about.

 

Walton

Mechanics & Punctuation, free, 20 page guide to everything punctuation  Build Your Book,  free 98 page guide to designing your book  CS Digital  understand CS digital possibilities GIMP, free, tutorials, GIMP, GIMP Help, excerpts from GIMP Supremacy Supremacy  Contact for graphics, design, and typesetting help  Disclaimer: all statements of apparent fact in this post are empirical inferences based on observational data. These are idiosyncratic in nature and have not necessarily been subject to verification.

Level 5 9,257 posts since
Aug 22, 2008
Currently Being Moderated
25. Jul 1, 2011 9:36 AM in response to: troffer
Re: Have you printed Color Photographs in your book?

troffer wrote:

In the case of supposed-to-be-red-but-is-sometimes-pink-and-sometimes-brown when there have been no changes in the source file, what does cause this obvious color shift other than printing techology issues or excessive variation in the printing process?

I don't know.  My guess was that the biggest problems started at the source (e.g., using something like Paint to create the cover design and a freebie tool with no settings options to create the PDF). 

 

Beyond that, in most digital workflows the images in a PDF/X-1a compliant file are passed through the RIP to the imagesetter -- whereas with a PDF file that contains RGB images, the images have to be converted to CMYK on-the-fly at the RIP.  So the "different day, different press, different paper, different result" scenario seemed plausible.

 

Further, the settings that were often being recommended here for PDF conversion (High Quality Print and Press Quality) allow/retain transparency (which CS would flatten, potentially causing color shifts -- they issue a warning message about that now, but didn't back then). 

 

Finally, it didn't appear (to me) that the CS workflow was honoring embedded RGB color profiles, which added another layer of uncertainty as to what would happen in print.

 

Restating something I've mentioned in other posts:  I've never suggested that a self-publishing author convert RGB images to CMYK.  The thing that usually generates these exchanges are (from my perspective) declarations regarding CS's workflow and processes that are made as though CS has officially told us all that or published it in manual somewhere -- when they haven't really told us squat, and when my experiences have been incongruent with whatever is being declared.  I think it's a disservice for us to speak on behalf of CS and post things like, "CS does . . .", "CS uses . . ." as though it's a known fact and common knowledge, and to advise new members who have design and print experience and who possess the right tools to produce output designed for commercial print that all that stuff is more or less useless if you're printing through CS.  It isn't.

Level 5 14,225 posts since
Sep 5, 2009
Currently Being Moderated
26. Jul 1, 2011 1:15 PM in response to: Lighthouse24
Re: Have you printed Color Photographs in your book?

I think it's a disservice for us to speak on behalf of CS and post  things like, "CS does . . .", "CS uses . . ." as though it's a known  fact and common knowledge, and to advise new members who have design and  print experience and who possess the right tools to produce output  designed for commercial print that all that stuff is more or less  useless if you're printing through CS.  It isn't.

 

I know you were addressing toffer, not me, but I couldn't help but overhear. There are similar to comments you made my way the other day, so at the risk of being overly sensitive, they do seem to cast a shadow in my direction, and I take umbrage.

  • I am not employed by nor affiliated with CS.  I can no more speak for them than any CS member can.  I have never suggested, postured, presumed to be more than or other than a CS member--speaking in behalf of himself only.
  • Anything we say, in any capacity, about anything, to anyone, is, ultimately (thinking of the skeptics, or various epistemological theories)  an opinion. To exaggerate (I said that the other day, and you missed it, so I'll repeat: I am about to use hyperbole): Earth's magnetic North is congruent with what we call the North Pole and it will be there tomorrow.  According to current research, that is an opinion and an assumption: the poles have, according to new research, reversed before, and could tomorrow.  Do I need to say, "In my opinion, the North Pole will still be within the Arctic Circle tomorrow"?
  • In the spirit of your previous complaints that I am making statements of apparent fact that have not been independently verified, I did add this disclaimer to my signature (I guess you didn't see it, or you wouldn't have needed to write that post to toffer):

Disclaimer: all statements of apparent fact in this post are  empirical inferences based on observational data. These are  idiosyncratic in nature and have not necessarily been subject to  verification.

Although the style touches on the purple for effect, that is a fair and accurate statement that precisely addresses your complaint. Humor should not discredit logic.

  • In CS Digital, page 3, I have written:

     The conclusions, pages 90-94, are based on my observations—qualitative and subjective, not quantitative and objective. The tests and demonstrations may or may not tell the same story to you, but with CS Digital, your decisions will be based on what you observe, not what you read online or guess. Proof everything: all images are different and our definitions of “good printing” probably vary from person to person.

      Perhaps most important for CS Digital to be an effective working tool, download the test images used on pages 44-45 and 48-49 from http://www.12on14.com/pages/createspace.htm

      Please let me know if you experience something different from what is in this book.

I have paraphrased that in at least one of my posts.

  • I have tested several dozen graphics issues involving art preparation and how CS prints from that art. I may have started from an experience with what CS and BookSurge could print going back to 2002, but I tried to design my tests to demonstrate things today; further, I designed them to demonstrate what we do not know as fact from CS contained in printing guides, facts, etc. written by them, but could conclude--through sound, cogent and verifiable reasoning and observation. I did not set out to prove anything. I set out to discover. 
  • Fact:
    • the inks used by CS to print CS Digital are not traditional offset inks;
    • the ink colors used by CS to print CS Digital are perceptually different form traditional (Pantone-like, PMS, etc.) off set process colors;
    • in printing CS Digital CS used at least two different halftone screens based on counting dots per inch, referred to as lpi: 124lpi and 174 lpi (these are my counts, in the case of 174, an expert referred to this as "175 lpi" I presume that is still in the ballpark of the "factual");
    • counting dots to establish AM/XM halftone line frequency of a printed piece not printed by them is how every expert I know, with one exception, does it (it is important to note that FM dots are different, but in the examples that I have examined, CS did not use them);
    • in printing CS Digital CS used color profiles (related to both RGB and CMYK color spaces) that are different from what I have in my edition of Photoshop;
    • CS stripped away some or all color profiles in CS Digital and replaced them;
    • CS printed CS Digital in Black and White, and the screen frequency that I counted was 76 lpi. 

 

Those are facts. (and, there are more). They are not opinions. They relate to things that CS has done  in printing CS Digital. There is no logical reason (not for soundness or cogency) to qualify them--certainly not when I often, not always, do in fact offer a qualifier that refers to tests or my book.  I have also sought verification from other people . . . oops, I guess my disclaimer is at least partially false.

 

I will acknowledge here and now that I may have made a category error, specifically the fallacy of composition: I have drawn the conclusion from my tests that CS prints a certain way, form one small set of tests, when, of course, I could have had, by shear luck, the same press with the same operator, same batch of ink, same batch of paper.  Perhaps I need to consider that possibility more, it has the hint of possibility.

 

(I'm having a hard time here, I know I should be speaking in the passive voice, and using the universal and rhetorical "we," but I just can't.)

 

This really is all such bs in it's rankest form.  I do apologize to the community for being "silly."  Gosh, after agreeing to run a test at your request, and having admitted I was probably wrong about something yesterday . . . . there it is . . . . I am reminded that in many cultures, admitting an error or apologizing is a sign of weakness that virtually demands attack.   Silly me, now I'm in a muddle over cross-cultural perspectives.

 

If any new member, with or without commercial experience, who wants clarification on my experience gaff cited above, contact me, I'll try to explain, however, please know in advance that my conclusions are based on observational data, etc., etc., etc. I will also try from now on to qualify my statements regarding Word, InDesign, early music, Photoshop, visual arts, printing, scanning, photography . . . Fair enough?  (But wait, didn't you chastise me for referring to my experience? . . . . )

 

W

Level 3 573 posts since
Sep 24, 2009
Currently Being Moderated
27. Jul 1, 2011 7:59 PM in response to: Lighthouse24
Re: Have you printed Color Photographs in your book?

Wow, I'm feeling a bit like the stooge in the middle right now (although "Curly Joe" had more hair, as I recall).

 

Just to clarify: I don't think I've ever said much of anything about what "CS does" or the equipment that "CS uses" (it doesn't appear that you said I did, but I just wanted to make that clear).

 

And I don't really care too much, either, as long as the output looks like it is supposed to.

 

For a standard input, a standard process should produce a standard output within an acceptably small range of normal variation.

 

Unfortunately, that doesn't seem to be the case here, because, as you indicated, we don't have much of anything for input standards. And, summarizing from some other threads, it also seems that there may be several different undefined, non-standard processes; and that the output seems to vary accordingly and, apparently, with some very noticeable randomness.  In my car, when I turn the top of the steering wheel to the right I expect the car to turn right---every time---not sort-of to the right, nor for the radio to change stations one time and the windows to roll down the next.

 

I just finished a book about the Apollo space program, and it describes how the people who designed the Command Module made it so that when the stick was pulled back, the nose went down instead of up.  Nobody had much of a problem with that except for the people who knew how to fly and were going to be flying the thing on a long trip.  It got standardized.

 

In my view, it would be beneficial for CS to provide input specifications having sufficient detail to allow us to obtain predictable results without pulling out what's left of Curly's hair, and processes stable and predictable enough to deliver them.

Level 5 14,225 posts since
Sep 5, 2009
Currently Being Moderated
28. Jul 2, 2011 8:17 AM in response to: troffer
Re: Have you printed Color Photographs in your book?

I'm sorry if you feel like Curly . . . better than feeling like Shemp (my opinion).

 

I did not respond to your question, nor did I direct my response to you . . . so you're in the middle of only one side, which gives you a wide open escape route.

 

I hope you didn't take anything I said as if it were directed at you, it wasn't.

 

As for your question:

 

I don't think "red" should print pink or brown. If it does (if your PDF shows red), then I'd contact Support.

 

Small shifs in color occur when RGB, which has a big footprint, is converted in Photoshop to CMYK or by most printers. If all color were represented by a placemat, RGB would be a dinner plate, and CMYK would be a salad plate.  Colors outside of CMYK are out-of-gamut.  There are four common ways for the RGB > CMYK conversion to occur in an attempt to deal with the those can't-print colors.  For the most part the shifts are small: most notable in either the loss of excitement, or in flesh tones.

 

Usually the out-of-gamut colors in RGB are the saturated ones.  In CS Digital I show several optical illusions to point out that even when there are shifts or colors are clipped, what we see (not the original artist) is an image that stands within it's own color world. We can have rather large shifts or compression without thinking anything is out of place.  Example: most people would think that Ansel Adams (just to pick a name) did a darn good job capturing the drama and luminence range of a scene.  In some of his landscapes the luminence range would have been 10,000 lumens outside.  His film had a luminence range of  200.  His paper had a luminence range of 100.  (These are numbers I read and am recalling from an article many years ago.)  So it's not the actual range we see, not even close, but it is how he handled representing that range.

 

Take a look at the Munker Illusion: http://www.michaelbach.de/ot/col_Munker/index.html. People who rely soley on colorimetric measurement would have no problem with this--3 colors are 3 colors; the green is the specified green. Period.  But if you wanted/needed the greens to actually look alike, you would have to print 4, not 3 colors!  Hence, I often quality things with the word "perceptual."

 

I often work in both RGB and CMYK (and LAB) for my own work. In CMYK I can easily adjust the Yellow and Black.  I figure what I lose in colors being clipped or converted, I gain in control.

 

What Indigo does is to use non-traditional ink colors and it uses color profiles, which are proprietary--not standard with Photoshop, that permit a dead-on perceptual match.  And in my tests show, that CS prints best from RGB, because, well, it prints best from RGB. Which means for most CS member (who aren't reading this), working with their cameras and scanners, putting totegther covers and illustrations, etc., they are happy.  People like me, would do what they have done for 10-40 years (or were taught at RISD or RIT) and their work won't be represented well, and they won't be happy. And they won't know why.  (Ooops . . . too much opinion here.)

 

Walton

Mechanics & Punctuation, free, 20 page guide to everything punctuation  Build Your Book,  free 98 page guide to designing your book  CS Digital  understand CS digital possibilities GIMP, free, tutorials, GIMP, GIMP Help, excerpts from GIMP Supremacy Supremacy  Contact for graphics, design, and typesetting help  Disclaimer: all statements of apparent fact in this post are empirical inferences based on observational data. These are idiosyncratic in nature and have not necessarily been subject to verification.

Level 3 573 posts since
Sep 24, 2009
Currently Being Moderated
29. Jul 2, 2011 4:18 PM in response to: walton
Re: Have you printed Color Photographs in your book?

No problem.

 

Conversion from one color space to another is just straightforward matrix arithmetic as long as the chromaticity of each of the primaries and the white point (relative amplitude of each primary required to make a neutral gray) are known.  Since it is just arithmetic, we should get the same answer every time.  If we don't, something is wrong.  Even if we use a fancy many-primary color system, the conversion is still just arithmetic and we should get the same answer (color) every time.  That color should be reproduced the same every time within the range of possible solutions (gamut) of the color system.

 

I agree that out-of-gamut colors cannot be accurately represented.  It is also not really possible to reproduce the gamut of a color space that uses additive primaries (RGB) in a color space that uses subtractive primaries (CMYK or other print medium) regardless of the choice and number of primaries.

 

Different RGB color spaces are not really the same, either.  The HDTV gamut is larger than the SMPTE gamut, for example, so HDTV colors don't transfer fully to SMPTE, but SMTPE colors will reproduce well with HDTV primaries provided the color space is rotated and the white point shifted to map one color space to the other as well as possible.  (An unmodified HDTV image has a greenish cast when displayed on a SMPTE system).  The translation is still just arithmetic and can be done on-the-fly by any digital system, although most don't.

 

In my vague recollection, Adams measured the technical characteristics of every part of his system and knew the curves and accuracy of each lens, batch of film, printing paper and related processes, including process modifications to predictably alter them.  He measured each important part of his composition to know where it would fall and whether process modifications would be beneficial.  Using this technical mastery, he was able to map one contrast range (original scene) to another (film) to another (paper) and incorporate his vision, interpretation, and emphasis in a predictable and repeatable way that allowed his production prints to be very much alike, even over time.  Zone V was Zone V every time (subject to interpretive needs).   He was doing many years ago an analog grayscale version of what I described above.  (I seem to recall that he was such a detail nut that he carried around a spare rear axle for his station wagon just in case it broke.)

 

So, for any claim that says 2 + 2 != 4 any more, I propose that the math isn't being done correctly.

 

(Unfortunately, that Munker link requires Flash so I didn't view it.)

Actions

More Like This

  • Retrieving data ...