Professional Web Applications Themes

Color Theory Questions - Adobe Photoshop 7, CS, CS2 & CS3

Greetings all, I haven't read this forum in quite some time, but if memory serves me correctly, there are quite a few users here that have a great deal of knowledge of color theory. I work for a rotogravure printing company and am in the process of establishing correct CMYK engraving curves. In essence, I have the abillity to print a 100 step tone scale (0-100%) where 50% can be anything I want (i.e. I can make it so that 50% prints 75% or 20%). My research has shown that traditionally, 50% isn't 50% - more like 60-70% which makes ...

  1. #1

    Default Color Theory Questions

    Greetings all,

    I haven't read this forum in quite some time, but if memory serves me correctly, there are quite a few users here that have a great deal of knowledge of color theory.

    I work for a rotogravure printing company and am in the process of establishing correct CMYK engraving curves. In essence, I have the abillity to print a 100 step tone scale (0-100%) where 50% can be anything I want (i.e. I can make it so that 50% prints 75% or 20%). My research has shown that traditionally, 50% isn't 50% - more like 60-70% which makes absolutely no sense to me. Why, as a printer, would I not want my dot percentages to be as linear as possible (50%=50%, 75% = 75%, etc.)

    It makes logical sense to me that an image would gain the most benefit by having as linear a dot area reproduction as possible. My current goal is to reproduce a tone scale created using Kodak ogue Matchprint material. We have output film with a totally linear scale, then expose the material per supplier specs. I can adjust the image setter so that the matchprint material reads virtually linear - yet if a truly linear film is used, the material has a curve to it (50% ain't 50%).

    So, color theory experts out there in the world, what is the general consensus when reproducing images via CMYK - I would imagine the truer the dot percentages are, the more accurate the CMYK print will be to the initial digital image being duplicated (yes, you can create curves in PS, but most designers don't do this, leading to a linear image on screen).

    Thanks for all your help. Guest

  2. #2

    Default Re: Color Theory Questions

    Hi Stephen,

    I'm not really inside colour theory, but the non-linearity might be needed because ink tends to "bleed" into the white paper surrounding the dot.
    That way I would imagine that 50% dot coverage (intended) actually produces a visible coverage of maybe 60% or so.
    Then a non-linear offset (or bias) could prevent this.

    Rob Guest

  3. #3

    Default Re: Color Theory Questions

    Since I am getting older, my memory may be suffering, but if I'm right, a
    50% dot means that 50% of the dot pattern is opaque, the other 50% is clear.
    This effectively cuts the light transmission in half. If we carry this over
    to photographic density, this is a density of 0.30 - that being the common
    log of the reciprocal of opacity (2). A photographic density of 0.30
    attenuates one half of the light. Photographic density is stated in log
    values whereas dot percentage sticks with a linear scale.

    My background is in photography, but I served a few years in color
    separation. I remember having problems with the same question you pose, but
    the above is what I remember. This probably doesn't fully (or even
    partially) answer your question, but maybe it will give you a start.

    Cooter Guest

  4. #4

    Default Re: Color Theory Questions

    You're exactly correct, Rob. I good example of this would be the difference in printing on newsprint as opposed to film.

    I can make a 50% "block" print any dot percentage I want with the use of gamma curves (what Rob is calling an offset or bias). It is very possible for me to print a block that reads 50% with a reflective densitometer. Yet, many "experts" are saying that we shouldn't be printing 50% as a true 50% - which makes no sense to me. This is why I'm shouting out to all those color theorists out there. Guest

  5. #5

    Default Re: Color Theory Questions

    Yes, but IMHO there would be no rule for it.
    I like to write with a fountain pen, you know, with liquid ink, and sometimes I have paper that has a very high "bleed factor": when I turn the page, side B appears to be unusable because it all bled through.

    The point is that it sometimes is hard to predict what offset one must use. Or bias. (I'm a broadcast eng, so pardon the wrong jargon).

    The problem for printers like you is perhaps bigger if you judge the quality of prepress on a CRT monitor: they also bleed but inverse!
    (white to colour. Rats! life aint easy).

    I don't exclude that I'm wrong, but this is my two (euro)-cents.
    Rob Guest

  6. #6

    Default Re: Color Theory Questions

    The problem is, if you compensate for dot gain and make proofs from compensated film using an og proofing system that does not simulate the same amount of dot gain as the actual press, then what good is it? You have lighter CMYK numbers in the film making lighter proofs.

    The ideal, in the pre-computer days, was to proof un-compensated film, then when approval was given, create new film for plating that includes compensation. But nobody does that anymore...

    Now the best bet is to send uncompensated data to a good digital proofer, then send compensated data to the imagesetter film or plate. Guest

  7. #7

    Default Re: Color Theory Questions

    Actually, the biggest problem is the fact that not many people understand gravure - and the majority of proofing systems - og or digital - are designed for offset/flexo. The actual dot created by an electromechanically engraved cylinder is totally different than the dot created by CTP systems or traditional plate exposure. Please be aware that I am not attempting to compensate for dot gain at all - I am mearly posing the question as to what one should really expect to print in terms of dot area - should 50% be 50% or should it be something different? If it should be something different, then why fool ourselves and call it 50%. Guest

  8. #8

    Default Re: Color Theory Questions

    Ah, so it's a philosophical discussion...

    Rotogravure dots are kind of like little diamond shaped intaglio wells aren't they?

    What would a true 50% rotogravure dot be? Guest

  9. #9

    Default Re: Color Theory Questions

    Does gravure actually have dots? I was under the impression that gravure is og in that the deeper the color the heavier the ink. The "dots" are simply a mesh that separates the adjacent tones.

    How does one prepare a gravure with a digital process?

    One other thing. Middle gray as represented by the lightmeter for correct exposure is actually 18% gray, not 50%. Guest

  10. #10

    Default Re: Color Theory Questions

    <> Guest

  11. #11

    Default Re: Color Theory Questions


    interesting doc.
    Isnīt rotogravure the same as copper gravure, here by cylinders instead of plates ?

    "A Rotogravure print is velvety - often richly so - for the same reason that a mezzotint has this delightful quality:
    it is loaded with ink, which is raised, and there are no sharp, wiry lines or edges. In certain places, under a
    glass, one can distinguish the regular pattern of a halftone screen, with this difference, however, that whereas in
    the ordinary halftone print (relief) the round dots of ink are seen to vary in size, but to be of uniform blackness,
    those of a Rotogravure print (intaglio) are square and of uniform size but vary in blackness. In many places the
    lines of the screen are entirely obliterated by the generous supply of ink. "

    How can the dots be square ? How can they vary in blackness ?
    Perhaps the dot size is a result of the micro surface and the etching process. Variations in blackness may be
    a result of a very thin ink layer which is torn when applied to the paper.

    Best regards --Gernot Hoffmann Guest

  12. #12

    Default Re: Color Theory Questions

    Actually, I believe there are only a few companies in the US that etch cylinders for gravure (if any at all). In gravure, a copper cylinder is etched or engraved with thousands of cells - these cells carry ink to the substrate. Practically all gravure printers in the US use equipment invented and sold by a company formerly known as Ohio Electronic Engraving(ers) or OEE for short. OEE was bought by Max Daetwyler (a major supplier of gravure equipment worldwide). Quite honestly, gravure was digital quite a bit before offset/flexo, due to OEE equipment.

    Traditionally, stepped films were wrapped around the copper cylinder after coating the cylinder with photoresist. The cylinder was exposed, then etched in the unexposed areas. The cylinder is then chrome plated for durability.

    In an OEE system, the copper is engraved by a diamond stylus. Voltage is passed through the engraving head, causing the stylus to "hit" the copper and carve a small cell. This stylus works at speeds of up to 8100 cells per second (latest and greatest technology) - older engravers run around 3000 cells per second. Off the top of my head, a cell is approximately 120 microns wide - depending on line screen.

    What many people don't understand, is that due to this engraving process, you can't rotate the angle of your dot. The diamond hits the cylinder the same way and spirals around the cylinder while it engraves the image. A miore is avoided by changing the compressions and screens of each CMYK color - compressions esentially change the shape of the cell - a 30 compression cell resembles a nice square "dot" while a 60 compression cell is very elongated. A typical CMYK setup for OEE engraved cylinders would look like this: Cyan - 175/60, Magenta 175/30, Yellow 174/45, and Black ~210/30-34 (keep in mind those second numbers are compressions, not screen angle).

    From a digital standpoint - 1 up files are sent to a programming station, where the files are RIPped, stepped, etc. Then that file drives the engraving head as it engraves the image into the copper cylinder.

    Building a proper gamme curve to ensure a nice tone from 100%-0% is accomplished by creating an engraving tone for each of the CYMK colors. This tone is built from Max engraving voltage to Min voltage (lower the voltage, the fewer dots are created). These tones are then printed with X substrate and Y inks. Each voltage value is read with a reflective densitometer and viola - a curve is built. So now we know that X voltage will produce Y dot area. Our current programing software allows 255 voltage steps - and a voltage index of 127.5 does not produce a dot area of 50% - it's generally around 35-40% depending on which color (different colors have different engraving specs, remember?)

    So back to my original question - wouldn't you want that final 100 step tone scale to print as linear as possible? Or is there some color theory consensus that images look better when 50% is printed at 60%, etc.

    If you made it this far, I hope you enjoyed the read ;) Guest

  13. #13

    Default Re: Color Theory Questions

    Also, if anyone would like to view some nice examples of printing with this process - about all major cigarette labels are printed in this fashion. If you were to purchase a soft pack of Marlboro Lights here in the US, I can almost guarantee is was made by the company I work for. Pall Mall, GPC, Basic, and Merit are a couple of brands off the top of my head that have tones on the labels - grab a label and a 10-30x loupe and enjoy. Guest

  14. #14

    Default Re: Color Theory Questions

    Great read, Stephen. I thought gravure had hit the skids a long time ago.

    I have an Edward Weston book printed in gravure. One of my most treasured.

    The etching process is similar to the mastering process on the old vinyl recordings. There, a diamond stylus engraves the information on a copper plate coated with acetate. What is interesting is if we got the stylus set too deep and actually hit the copper, kiss the stylus good-by! You are actually etching the copper. What is the geometry of the tip? Does it wear quickly?

    I expect that the curve you plot, voltage against density, is the voltage required to etch to a certain depth. So, I don't understand what you mean by "tone" in this discussion, as tone is usually a function of incident vs reflected light. A trade term?
    Also, at first you indicate that the number of dots is a function of the applied voltage, then later it's the area of the dot as a function of voltage. Now, if the ink varies in depth, I would expect a relationship between voltage (or power) and dot depth. So we seemingly have three different parameters all a function of applied voltage.

    A bit confusing to me.

    As for linearity, the eye is non-linear, so I would expect a power curve and not a linear curve.


    Lawrence Guest

  15. #15

    Default Re: Color Theory Questions

    As far as wear, I can't be sure, as I'm not directly involved in the day to day operations. The geometry of the stylus ranges from 100 degrees to 130 degrees (lower the angle, the more volume of ink the cell will hold). Generally speaking, I can get the same amount of volume from a 200 line screen engraving with a 120 degree stylus as I can a 175 line screen with a 130 stylus - that's really generalized though.

    I dunno, explaining myself via writing is not my strong suit. Basically, the higher the voltage, the deeper the diamond "chisels" out copper. A cell opening of about 25 microns is about is small as you can go and still produce a "dot" on 60# C1S fine quality paper. When I say "tone" I'm referring to a vignette (from 100% - 0%) created by half tone dots - 50% dot area meaning half substrate, half ink. Lower voltage produces a smaller cell, and fewer cells in a given area - thus once the ink is on paper, one can read this pattern with a reflective densitometer at get a dot area reading.

    What do you mean by "power-curve"? Guest

  16. #16

    Default Re: Color Theory Questions

    Ok, I get it. Going deeper means going wider as well.

    Sounds like an industry I would like. I enjoyed cutting masters. And, I dearly love the mezzotint look. My prints are now almost exclusively on a matte paper with K in the CMYK printer settings raised a couple of notches. I get a mezzo-like effect that way.

    Also, the pitting of the copper sounds very similar to the pitting the laser does in CD burning. Has any laser cutting tools become available to gravure printing? It would seem that the laser would really have great control of the geometry. And, no stylus' to replace!

    The cutting stylus for records is a complex shape having both cutting and burnishing facets to it. It is actually a chisel.

    A power curve is the curve produced by raising a number to a power, like y=x ^2, or y=x squared. Guest

  17. #17

    Default Re: Color Theory Questions

    Oops, double post!
    Sorry! Guest

  18. #18

    Default Re: Color Theory Questions

    There are several laser technologies, but the technology hasn't quite caught on in the industry. Traditionally, the copper cylinder is plated with chrome - making it extremely durable - and once the chrome is worn, it can be stripped and re-plated. The current laser engravers engrave into either a nickle or zinc (which one, I can't seem to recall at this moment). You can't remove chrome from these, so once it's worn, you have to re-engrave. Guest

  19. #19

    Default Re: Color Theory Questions

    Side note - gravure is still quite popular for bank notes, contracts, and official doents where you want to make forgery as difficult as possible. Guest

  20. #20

    Default Re: Color Theory Questions

    Gernot had asked "How can they vary in blackness?"

    Would that not be due to the depth of the cells themselves Stephen? Deeper cells hold more ink and make blacker dots? Meaning that a cell for a highlight is just a wide as a cell for a shadow, but since it is so shallow, it hardly holds any ink at all... I find this a bit hard to truly grasp.

    ....or maybe not, if you compare it to another intaglio process, fine art etching in zinc plates. Given an aquatint of a given coness (ogous to cell frequency) the longer time in the acid etches deeper wells for the ink... Yeah, that's it. Unh Hunh. Guest

Similar Threads

  1. graph theory
    By jimcath in forum Mac Programming
    Replies: 5
    Last Post: March 3rd, 10:46 AM
  2. string theory
    By klcmx in forum Macromedia Flash Actionscript
    Replies: 2
    Last Post: January 19th, 01:05 PM
  3. Source for OS Theory
    By Marmagya in forum UNIX Programming
    Replies: 7
    Last Post: September 28th, 05:34 PM
  4. theory vs practice ceases power
    By Xah Lee in forum PERL Miscellaneous
    Replies: 1
    Last Post: July 22nd, 06:57 PM


Posting Permissions

  • You may not post new threads
  • You may post replies
  • You may not post attachments
  • You may not edit your posts

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139