From: INTERNET:gowens01@coin.org, INTERNET:gowens01@coin.org Date: Thu, Dec 28, 2000, 7:13 PM RE: [colortheory] Color settings and upgrades I'll start with a question. I just upgraded my Mac to OS 8.1 (I've hesitated upgrading to 8.6 and 9.0 because I read that there were conflicts with Photoshop). I have both 5.0 and 5.5 in the computer and have set them to emulate the color settings for Photoshop 4.0. If I calibrate my monitor with Colorsync will I have a conflict with my photoshop settings? I've just read Mr. Margulis article in the December 2000 issue of Electronic Publishing about upgrades. I would like to upgrade to photoshop 6.0 but as I mentioned before I'm still operating on 8.1 and if I would upgrade to 9.0 come January it will be time to upgrade to OS X. I'm running a 7100/80 Mac. I will have to buy a new computer to use OS X. And if I upgrade to Photshop 6.0 it will not be native to OS X. And Photoshop hasn't said anything about a native upgrade for OS X. Thank you for your articles in Electronic Publishing Mr. Margulis. I did follow your instructions to get Photoshop 5.0 to emulate the 4.0 color setting. Gary Owens ------------------------------------------------------------------------------- From: Dan Margulis, INTERNET:76270.1033@compuserve.com Date: Thu, Dec 28, 2000, 9:16 PM RE: [colortheory] Color settings and upgrades Gary writes, >>I just upgraded my Mac to OS 8.1 (I've hesitated upgrading to 8.6 and 9.0 because I read that there were conflicts with Photoshop). I have both 5.0 and 5.5 in the computer and have set them to emulate the color settings for Photoshop 4.0. If I calibrate my monitor with Colorsync will I have a conflict with my photoshop settings?>> No, but your images won't look the way they used to on screen, if that upsets you. That's true for PS 4 as well as PS 5. The program's calculations aren't affected by what you may do to calibrate the monitor. >>I've just read Mr. Margulis article in the December 2000 issue of Electronic Publishing about upgrades. I would like to upgrade to photoshop 6.0 but as I mentioned before I'm still operating on 8.1 and if I would upgrade to 9.0 come January it will be time to upgrade to OS X. I'm running a 7100/80 Mac.>> I think with a box that old your decision to avoid 8.5 and up is a smart one. That rules out the use of Photoshop 6. When you buy a new Mac, then you'll have a new OS, and you can decide where to take it from there. Dan Margulis ---------------------------------------------------------------------_-> From: James Steincamp, INTERNET:steiny@cavtel.net Date: Thu, Dec 28, 2000, 9:36 PM RE: Re: [ColorTheory] [colortheory] Color settings and upgrades Hi Gary... > I'll start with a question. I just upgraded my Mac to OS 8.1 (I've hesitated > upgrading to 8.6 and 9.0 because I read that there were conflicts with > Photoshop). I have both 5.0 and 5.5 in the computer and have set them to > emulate the color settings for Photoshop 4.0. If I calibrate my monitor with > Colorsync will I have a conflict with my photoshop settings? I'm not sure where you read that there were conflicts or what the nature of the conflicts happen to be... but I've been using OS 9 for long enough that I have forgotten what it's like to use previous versions of the OS. I also haven't experienced any problems... outside the rude awakening that Photoshop 5 dealt early adopters a while back. How you choose to configure Photoshop's color setting is a major source of religious schism in the realm of graphic arts; pitting the legalists against the mystics. Enter at your own risk! ;p Your question about calibrating the monitor (I'm assuming it's a ColorSync monitor) requires that your wrap your mind around the peculiar way that Photoshop (and Managed Color) has evolved. Here's a mini 'Guide for the Perplexed': Assuming that you are working with RGB, it is necessary to understand that Photoshop's 'Working RGB Space' [Adobe RGB (1998) if you choose the US Prepress Defaults or sRGB if you accept Photoshop's defaults] has nothing to do with the fact that your monitor happens to work in RGB... or for that matter any RGB device (scanner, etc.). This is what Managed Color folks call 'device independent color.' Basically a Working RGB space is an abstract model for containing your RGB file and its color information. Again it has nothing to do with your monitor... and, in fact, it is almost irrelevant... except to understand that the relative 'size' (that's 'Gamut' in techno-speak) of your Working RGB Space can stretch or compress your files... essentially clipping the color short (in the case of sRGB when used for prepress) or leaving significant areas of color outside of what a device is capable of reproducing (as would be the case if you used Adobe RGB to prepare your web images for the average surfer with a cheap PC monitor). My advice is: use Adobe RGB if you are preparing files for print and if your a web guy use sRGB as your working space. Other spaces have specific uses... but that's a topic for another day. Then there's your monitor. You monitor is an RGB device and it behaves a bit differently over time with colors shifting about and so on. Many folks will calibrate their monitor to the same standard every day, or two weeks or whatever to attempt to control this color drift. All it means to calibrate is to set the device in question to behave according to a know standard: a specific color temperature, gamma and brightness. That's it. A lot of folks like to do this... even non-Managed Color folks. Then there is profiling... which in a Managed Color environment typically means to create a ColorSync (or ICC) profile. Profiling is an acknowledgement that no device is perfect (even when calibrated) and serves as an attempt to describe a device's (your monitor in this case) imperfections. ColorSynch then attempts to eliminate those imperfections from the process so that you are not changing your wonderful RGB file based upon incorrect assumptions. For example you may try to color correct in a room that has fluorescent lights... you're seeing a lot of green in that case... and correcting under those conditions would likely leave your file too magenta under tungsten lighting. A profile attempts to correct that situation. The catch is that a profile is only as good as the repeatability of the circumstances under which it was created. If your monitor drifts for months without calibration or the lighting in your viewing room changes... your profile is pretty much useless. Profiling is therefore used in conjunction with calibration. With that said... the answer to you question as to whether calibrating your monitor will conflict with Photoshop's color setting is 'No.' Why? It's simple: all that happens when you calibrate your monitor is that you change the brightness, contrast and color settings of your monitor. Nothing happens to your file... you're tinkering with the RGB guns in your monitor and (in some cases, with the RGB output of the video card). That's it. Profiling the monitor with the ColorSynch software will attempt the have you 'eyeball' the color temperature of your viewing room, as well as the gama and brightness of your monitor. In Photoshop 5.x you tell the application to take the profile you have created into account by selecting 'Display Using Monitor Compensation' under the RGB set-up dialogue. All that happens here is that Photoshop will attempt to compensate for the conditions of your monitor and viewing room by altering the monitor's output (color, brightness, etc.) according to the data in the profile. Again, as with calibrating, nothing is happening to the file itself... it's just about what you're being presented on screen. The only problem is that should you create a really whacked out profile, you will see truly whacked results on screen and 'correct' your file accordingly. The little doughnut shaped 'thingy' that ships with ColorSynch monitors is hardly a precise method of creating a profile... it's slightly better than useless. In your case you have chosen to ignore the color management features of Photoshop 5.x and you can continue to do that with Photoshop 6.0 if you wish. Your questions above lead me to believe that you would like to have an accurate presentation of you file on screen without coping with the Managed Color issue. If that is indeed the case, you have three options: 1) Learn to understand that (without significant time, cash and effort) what you see on the screen is a lie. This is what Mr. Margulis wants you to understand in his book... learn the numbers, feel the numbers. It takes a bit of practise but it works just fine. 2) Dive in and learn how to Manage Color for yourself. It takes a good bit of cash, time and effort... but it works just fine. 3) Spend a few bucks on calibration software and hardware to regularly calibrate and profile your monitor. At least in this case what you see on screen will remain consistent and you can learn the relationship between the screen and your prints. The disadvantage here is that you have not taken the time to learn the numbers, nor have you created a Managed Color Workflow as you have no data concerning your input (the scan) and output conditions (the print). It's a bit like photography... I used Nikon F2s for a long time and a hand-held incident meter, then learned how to read the numbers on the lens barrel for focusing. It required some trial and error, but in short time it worked very well. Then I got my new Nikon F100 last year... the manual was several hundred pages and I found that the 'features' of the camera made it difficult to use. It was more difficult to learn the new features than it was to understand the 'sunny f/16 rule.' But in time I learnt to use it and I have to say that when it comes to doing flash photos it sure beats calculating a guide number on the fly. > I've just read Mr. Margulis article in the December 2000 issue of Electronic > Publishing about upgrades. I would like to upgrade to photoshop 6.0 but as I > mentioned before I'm still operating on 8.1 and if I would upgrade to 9.0 come > January it will be time to upgrade to OS X. I'm running a 7100/80 Mac. I will > have to buy a new computer to use OS X. And if I upgrade to Photshop 6.0 it > will not be native to OS X. And Photoshop hasn't said anything about a native > upgrade for OS X. I wouldn't sweat OS X for a while. Our shop has been beta testing the new version of OS X Server (which is quite different than what was released two years ago) and I have to say that OS X is enough of a change that many Mac users will balk at the transition. The real power is under the hood and more likely to appeal to Linux and BSD geeks than the average Mac user. For the immediate future the power of OS X is an alternative to Windows NT and Linux as a server in a Mac environment... and an improvement over ASIP 6.x in terms of speed, stability, scalability and security. Considering that you have chosen to remain at MacOS 8.1, the changes to OS X are infinitely more fraught with complexity and reservations. Most folks are going to find that the best option is to wait for at least a year, several revisions and native application support (if it ever comes). In the meantime, watch and learn. Hope that was helpful... -james ------------------------------------------------------------------------------- From: Bob Smith, INTERNET:rmsmith@calpha.com Date: Thu, Dec 28, 2000, 11:35 PM RE: Re: [colortheory] Color settings and upgrades Chris Murphy wrote: > You want a new computer anyway. Anything you get today will come with Mac > OS 9, and run Photoshop 6 so much faster than the 7100 making it worth > it. Just for example... I've spent the last couple of days moving from an aging and heavily upgraded 7600 to a G4 with twin 450's. A 120MB CMYK layered Photoshop file opened in 1 minute, 40 seconds on the G3 powered 7600. It opens in a little under a minute on my 400mhz G3 Powerbook. It opens in 6 seconds on the G4. That's a pretty damn serious performance improvement for about $2500. Bob Smith ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Thu, Dec 28, 2000, 10:02 PM RE: Re: [ColorTheory] [colortheory] Color settings and upgrades >I'll start with a question. I just upgraded my Mac to OS 8.1 (I've >hesitated upgrading to 8.6 and 9.0 because I read that there were >conflicts with Photoshop). I'm running Photoshop 4, 5.5 and 6 on Mac OS 9.04 with no problems. > If >I calibrate my monitor with Colorsync will I have a conflict with my >photoshop settings? No. They don't interact. >I'm still operating on 8.1 >and if I would upgrade to 9.0 come January it will be time to upgrade >to OS X. If you just recently upgraded to 8.1, then you're probably working on a machine that won't run X anyway. And in any event, Mac OS X is vaporware at this point. It's a distance dot on the horizon. There will be no point to move to X until there are native versions of applications you use on a regular basis. We could get surprised and see a Carbon version of Photoshop 6 before summer but I highly doubt it. And while it will run in Classic on Mac OS X, I don't see the point. There is no benefit running Photoshop in Classic on Mac OS X. If it's going to crash on 9 it will still crash Classic on X. Mac OS X is something that is for really early adopters once the final version is released. I don't see it being a factor until about September of next year (or more accurately, six months after it's released). Be aware that Mac OS X will not come with Mac OS 9 (i.e. no Classic compatibility unless you already have Mac OS 9). So everyone is better off worrying about Mac OS 9 and upgrading to it before being even remotely interested in X. >I'm running a 7100/80 Mac. I will have to buy a new computer >to use OS X. And if I upgrade to Photshop 6.0 it will not be native >to OS X. And Photoshop hasn't said anything about a native upgrade >for OS X. You want a new computer anyway. Anything you get today will come with Mac OS 9, and run Photoshop 6 so much faster than the 7100 making it worth it. When it's time to consider moving to Mac OS X in about a year, the computer you buy today will be able to run it. Chris Murphy ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Fri, Dec 29, 2000, 12:27 PM RE: Re: [ColorTheory] [colortheory] Color settings and upgrades on 12/28/00 7:34 PM, James Steincamp at steiny@cavtel.net wrote: > Then there is profiling... which in a Managed Color environment typically > means to create a ColorSync (or ICC) profile. Profiling is an > acknowledgement that no device is perfect (even when calibrated) and serves > as an attempt to describe a device's (your monitor in this case) > imperfections. ColorSynch then attempts to eliminate those imperfections > from the process so that you are not changing your wonderful RGB file based > upon incorrect assumptions. You were doing GREAT up to this point. I take some issue with the above statement. Profiling is just a method of describing the color a device is producing at a given time (we'd hope that such a device would remain consistent to that profile description). There isn't anything specific about perfection of a device or the "elimination" of imperfections. The profile simply "tells it like it is" and that's quite useful for doing conversions from space to space (you can't get to point B without describing point A) and it's useful for conducting a correct preview of a file in Photoshop or other CMS savvy applications. ColorSync doesn't attempt to eliminate imperfections and in fact, ColorSync itself is simply a system level engine for dealing with our device profiles. Andrew Rodney ------------------------------------------------------------------------------- From: James Steincamp, INTERNET:steiny@cavtel.net Date: Fri, Dec 29, 2000, 7:52 PM RE: Re: [ColorTheory] Color settings and upgrades Perhaps I was a bit harsh in stating that a profile attempts to describe a device's inherent limitations, but I was more interested in making a philosophical point than a technical one. 'Telling it like it is' is only necessary because of the imperfections (ie differences) that each device brings to the workflow. As a practical example we have two Epson 3000 printers at work... and both, left to their own devices, will render the same image differently. Not just theoretically different mind you, it's apparent even at a glance. A good profile (in a proper workflow) would describe the difference between these two (theoretically identical) devices and then allow me to eliminate the influence (the imperfections, really) that my two renegade printers have on my files. The functionality of a Managed Workflow is to eliminate the impact that different color spaces (the various forms of RGB, CMYK, etc.) and device spaces (monitor, scanner, printer, press, etc.) have on the images that are processed through it. While you are absolutely correct that ColorSync & profiles do nothing on their own to eliminate imperfections, the point of setting up and maintaining a Managed Color workflow is precisely to eliminate imperfections. A profile only becomes truly useful when used in a workflow that remains both quantifiable and consistent. Many folks have neither the time, the patience nor the financial resources to make that happen. In addition, more than a few users choose to approach Managed Color with the incorrect assumption that it will free them from worrying about CMYK values -- with the same sense of false confidence that many amateur photographers have in their fully automated SLR cameras. I believe pretty firmly that, at this point in the development of Managed Color, the average user is far better off learning how to relate to the CMYK values and the basic concepts of color reproduction outlined in Mr. Margulis' book than they are delving into the constantly evolving methods of Managed Color. Not to say that we all won't be looking to learn Managed Color in the future, but with a firm grasp of the fundamentals we will be far better off when should we eventually decide that it offers an appropriate return on our investment. -james > on 12/28/00 7:34 PM, James Steincamp at steiny@cavtel.net wrote: > >> Then there is profiling... which in a Managed Color environment typically >> means to create a ColorSync (or ICC) profile. Profiling is an acknowledgement >> that no device is perfect (even when calibrated) and serves as an attempt to >> describe a device's (your monitor in this case) imperfections. > You were doing GREAT up to this point. I take some issue with the above > statement. Profiling is just a method of describing the color a device is > producing at a given time (we'd hope that such a device would remain > consistent to that profile description). There isn't anything specific about > perfection of a device or the "elimination" of imperfections. The profile > simply "tells it like it is" and that's quite useful for doing conversions > from space to space (you can't get to point B without describing point A) and > it's useful for conducting a correct preview of a file in Photoshop or other > CMS savvy applications. ColorSync doesn't attempt to eliminate imperfections > and in fact, ColorSync itself is simply a system level engine for dealing with > our device profiles. > > Andrew Rodney ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Sat, Dec 30, 2000, 12:30 PM RE: Re: [ColorTheory] Color settings and upgrades on 12/29/00 5:50 PM, James Steincamp at steiny@cavtel.net wrote: > 'Telling it like it is' is only necessary because of the imperfections (ie > differences) that each device brings to the workflow Differences yes, imperfections, not necessarily. Profiles don't correct, they describe. I make this distinction because in the last few years, profiles have received in some cases a bum rap and in other cases, they are simply misunderstood. Profiles are really just very robust labels that travel with a file and describe the color for the rest of the CMS. If you could care less about a CMS, the profile doesn't do a thing. > As a practical example > we have two Epson 3000 printers at work... and both, left to their own > devices, will render the same image differently. No surprise here. I can take two Fuji Pictrography 3000 which have an on board calibrator and calibrate each to the Fuji standard and get two different prints. Welcome to the world of device dependant color (where the color is dependant on the actual device). I can profile both Fuji printers and produce prints that do match. The profiles describe different conditions for each printer. A conversion into the RGB output space produces two files with different sets of numbers but prints that match. > A good profile (in a > proper workflow) would describe the difference between these two > (theoretically identical) devices and then allow me to eliminate the > influence (the imperfections, really) that my two renegade printers have on > my files. I'd agree with everything but the word imperfections. It's likely that if you send both devices the same neutral RGB file, you'll get two prints that are not neutral without a profile. One could say this is due to imperfections although both printers are imperfect. However, the files may be imperfect. We optimize the file to produce a desired result. The results are varying RGB numbers in output space to get to a goal. It's semantics of course. But by properly describing both the color we start with using a profile and converting that color into an output space using a 2nd profile (a descriptor of the device), we get what we want. So the profiles appear to correct a situation. What they really did was simply describe the color and thanks to that small feat, we get the color we expect. > A profile only becomes truly useful when used in a workflow that remains > both quantifiable and consistent. Absolutely. > Many folks have neither the time, the > patience nor the financial resources to make that happen. The alternative is? > In addition, more > than a few users choose to approach Managed Color with the incorrect > assumption that it will free them from worrying about CMYK values -- with > the same sense of false confidence that many amateur photographers have in > their fully automated SLR cameras. Not a bad analogy. With a fully automated SLR, as far as exposure goes, the exposure system assumes the world is 18% gray. That's a pretty good assumption in many cases. When you shoot back-lit stuff, snow, that black cat on the coal, the assumption falls apart. But for a vast majority of situations, it works pretty well. With a profile, if you've dialed in the CMYK conversion, the need for numbers is pretty low. There are situations where having info about highlight and shadow is useful and if you don't have a calibrated display, you can do some work using just the info palette. The numbers are great for correcting where you want to start at a certain point and move the color a certain amount in one direction or another. But there isn't much more use for those numbers. Both the auto SLR and the profile eliminate a lot of guess work in most situations but you can get hosed (more with the SLR then the profile). IF you understand the logic behind the auto exposure (18% gray assumption for all metering), and you can control the exposure system, you can always over-ride the bad assumptions based on this understanding. With a good profile, you get to the colorspace necessary for the device. You may still need to edit the file but the worst is over (as opposed to some generic or simply incorrect assumption about how to convert to CMYK). > believe pretty firmly that, at this point in the development of Managed > Color, the average user is far better off learning how to relate to the CMYK > values and the basic concepts of color reproduction outlined in Mr. > Margulis' book than they are delving into the constantly evolving methods of > Managed Color. I think that's a fantasy! I don't even buy that Margulis can do this all the time, with all CMYK output devices. It's clear that his understanding of CMYK to more standardized output (dare I say SWOP) is so high that he can make many valid assumptions about the numbers. But throw the files to anything that isn't a known fact to Dan and the system falls apart quickly. You just told me that the two Epson 3000's vary from just two models. Which assumption about the CMYK values based on what Dan would come up with (if he even had an intimate feel for this device) is correct? The numbers for one would produce a different result then the other right? Now ask Dan or someone with very well honed skills to provide you numbers for an Iris ink jet running Lyson archival inks to a rag coated paper that are correct. What about output to a Kodak Approval today but tomorrow Fuji ColorArt? And with this intimate info about numbers, take my CMYK file to that ColorArt and convert it to another CMYK space so I can proof it to my Epson 3000! Do you really think the "average" user can learn to do this? I don't. Now just ask the "average" user to simply have a profile that describes the devices he/she wishes to use. Getting from RGB to CMYK or CMYK to CMYK or whatever is really simply. Yes, getting that good profile is some work. The print shops should be supplying them no doubt. With the resistance to this mode of assisting customers, it's easy to blame the cost and complexity of managing color and say the system is faulty. Now take my fantasy where printers and service providers simply provide a profile they have made for their devices to their customers. It's not at all difficult to get the customer who can follow instructions to load the correct profile and get the job done in a vastly more controllable fashion. Yes, I've heard how difficult it is for printers and service providers to create profiles and "train" their customers to use them. Again, the alternative is what? Stick to the system we have now (the old pre Photoshop 5 color system) which is pretty chaotic and quite hit and miss. Both systems take work on the part of the service provider and the part of the user. Which will ultimately be more accurate, flexible, easy to use (provide nice stuff like a soft proof), cross rendering and file identification? I think the profile route clearly shows more promise in all cases. Getting people to implement this newer technology is a different topic all together. Andrew Rodney ------------------------------------------------------------------------------- From: James Steincamp, INTERNET:steiny@cavtel.net Date: Sat, Dec 30, 2000, 6:15 PM RE: Re: [ColorTheory] Color settings and upgrades >> 'Telling it like it is' is only necessary because of the imperfections (ie >> differences) that each device brings to the workflow > > Differences yes, imperfections, not necessarily. Profiles don't correct, they > describe. Understood... they describe the inherent limitations of a device. A perfect device would require no such description as it would have a large enough gamut to encompass the range of colors we are capable of perceiving and it would precisely match our working RGB color space; rendering other spaces unnecessary. But in reality we are left to cope with the imperfections of every element in the system... the cyan ink on press... humidity... my fickle Epsons... an old monitor and so on. > I make this distinction because in the last few years, profiles have received > in some cases a bum rap and in other cases, they are simply misunderstood. Managed Color has received a bum rap because a) in the beginning the marketing folks allowed the hype to exceed what was delivered ['Color is Child's Play' comes to mind or any product with 'EZ' in the name], and b) many of those who believed the hype thought they were getting an automatic, zero maintenance panacea for all their color problems. They failed to take into account the fact that a CMS requires rigorous discipline to keep the workflow under control (not at the service bureau or on press... where process control is the norm), but in house where that had never been a concern before. > Profiles are really just very robust labels that travel with a file and > describe the color for the rest of the CMS. If you could care less about a > CMS, the profile doesn't do a thing. This is actually one feature that is particularly useful. It's nice to know how someone else has chosen to process a file (usually, from my experience, louse it up) on opening... without making a call (which could, at least in my case, be half-way around the world in the middle of the night). > I'd agree with everything but the word imperfections. I think you object more to the moral connotation of the word than it's explicit meaning. ;P > It's likely that if you send both devices the same neutral RGB file, you'll > get two prints that are not neutral without a profile. One could say this is > due to imperfections although both printers are imperfect. However, the files > [themselves] may be imperfect. But we will never know precisely where the imperfection lies as we have neither a perfect device, nor a perfect instrument for measuring the imperfections. We are left to muddle through as best we can. > We optimize the file to produce a desired result. The results are varying > RGB numbers in output space to get to a goal. It's semantics of course. Or we deliberately distort the file in order to output as we would like. Here we begin to move into the realm of subjectivity. Which do you like better: Fujichrome's interpretation of color, or Kodachrome? What about selenium toning a fiber print to intensify the blacks and make the whites... err, blue (my wife swears that if you throw a blue shirt in with new whites the first time you wash them, you get a 'cleaner' white that lasts longer)? How many times has a client asked you to alter the color of some object in a print because they 'feel' that the grass should be greener or that the faces should be brighter. Scientifically they have no basis for making that call... or even, for that matter, to tweak the reproduction curve to give us a nice 's' shape and exaggerate the contrast at the toe and shoulder (at the expense of the mid-tones). Much of what 'color professionals' deal with on a daily basis is pure trickery... and unfortunately lies beyond the pale of what we can describe objectively. For that we have intuition, based on previous (and yes, imperfect) knowledge. >> A profile only becomes truly useful when used in a workflow that remains both >> quantifiable and consistent. > > Absolutely. > >> Many folks have neither the time, the patience nor the financial resources to >> make that happen. >> > The alternative is? Precisely what we have been doing without Color Management for years. Tinker an learn. >> In addition, more than a few users choose to approach Managed Color with the >> incorrect assumption that it will free them from worrying about CMYK values >> -- with the same sense of false confidence that many amateur photographers >> have in their fully automated SLR cameras. > > Both the auto SLR and the profile eliminate a lot of guess work in most > situations but you can get hosed (more with the SLR than the profile). IF > you understand the logic behind the auto exposure (18% gray assumption for > all metering), and you can control the exposure system, you can always > over-ride the bad assumptions based on this understanding. With a good > profile, you get to the colorspace necessary for the device. You may still > need to edit the file but the worst is over (as opposed to some generic or > simply incorrect assumption about how to convert to CMYK). Exactly... it is the fundamental knowledge about the process that allows the user to understand what the CMS is trying to accomplish and how to ensure that the workflow itself remains true to the desired endpoint. With a sound knowledge of CMYK and the weaknesses inherent in the system (all systems have weaknesses), the user is capable of grappling with the subjective issues that arise in the course of preparing images for output. >> believe pretty firmly that, at this point in the development of Managed >> Color, the average user is far better off learning how to relate to the CMYK >> values and the basic concepts of color reproduction outlined in Mr. Margulis' >> book than they are delving into the constantly evolving methods of Managed >> Color. > > I think that's a fantasy! The problem that I have with Managed Color at this point is that with each version of Photoshop and each incarnation of of ColorSynch and so on the user is left grappling with (mostly undocumented) issues like the differences between CMS engines and black point compensation as opposed to learning about 'the thing in itself,' color. ColorSynch, etc. has come a long ways in the last three years, but it will continue to change dramatically (for the better) and require a fair amount of effort on behalf of its users to keep up with the issues (ie. I really want to calibrate my Apple Cinema Display or how 'bout getting a reading off those new 'permanent' printers Epson released earlier this year... or what if the glaze on the surface of a round porcelain object wreaks havoc on my attempt to get a good reading off the decal?). > I don't even buy that Margulis can do this all the time, with all CMYK > output devices. It's clear that his understanding of CMYK to more > standardized output (dare I say SWOP) is so high that he can make many valid > assumptions about the numbers. But throw the files to anything that isn't a > known fact to Dan and the system falls apart quickly. Not really. If you remember in the early days of Managed Color there was the idea of profiling the press itself... a scientifically sound notion. But that fell by the wayside pretty quickly do to the cost involved... and now we aim for the all-important MatchPrint (or whatever). That's one step removed from the end product in the scenario that is most important to me. It's key to understand that all the digital precision and profiling information (and it's imperfections) ceases the moment the plates go on the press and that means I'm still left waiting for that call in the middle of the night to come down and do a press check... where I'm almost certain to want to tweak. There I am holding a contract proof saying... it looks like the shadows still have to much 'insert your color here' without the benefit of any scientific data, it's just me and what I understand about color. Your right that Mr. Margulis probably can't hit every 'print' on every device the first time... but neither can you (or anyone else for that matter). A CMS system is dependent on having valid data about the process in question; which means that you have to run a set of targets to provide the proper data for the CMS. That means, in essence, that you have taken two attempts to get to your finished result. I' reasonably confident that a competent numbers guy can hit any device (that behaves according to man's present understanding of color) with two tries. I can do that pretty well myself and I don't exactly have a decade's experience to guide me. > You just told me that the two Epson 3000's vary from just two models. Which > assumption about the CMYK values based on what Dan would come up with (if he > even had an intimate feel for this device) is correct? The numbers for one > would produce a different result than the other right? Naturally... or, to sharpern your point, as one SWOP press differs from another. The process is exactly as yours... run a test. A numbers guy will fall back on years of experience tweaking the numbers; a CMS guy will fall back on the knowledge he has learnt about profiles and workflow. > Now ask Dan or someone with very well honed skills to provide you numbers > for an Iris ink jet running Lyson archival inks to a rag coated paper that > are correct. What about output to a Kodak Approval today but tomorrow Fuji > ColorArt? And with this intimate info about numbers, take my CMYK file to > that ColorArt and convert it to another CMYK space so I can proof it to my > Epson 3000! Sure, the more you move away from a press and the CMYK we are all accustomed to, the better the case for moving to a CMS. Yet most folks are still focused on getting good color on press... where the bulk of money is currently being made. [Aside] One thing that I have learnt as an art director is that it can be particularly dangerous to hand the client an ink-jet proof... it allows that individual, with his or her grossly inadequate understanding of color, to think that they are now qualified to start doing my job. After all, 'I can make a "quick change," right?' In all honesty there is a degree of finality in showing the client a MatchPrint and the implicit understanding that more revisions will cost $65 - $100 a pop. ;) That's why the printers have resisted profiles... it puts the responsibility for managing color in the lap of an often clueless end user... who in turn will want to blame the printer when things go wrong. > Both systems take work on the part of the service provider and the part of > the user. Which will ultimately be more accurate, flexible, easy to use > (provide nice stuff like a soft proof), cross rendering and file > identification? I think the profile route clearly shows more promise in all > cases. Getting people to implement this newer technology is a different > topic all together. Pointing out a technology's weaknesses is different from dismissing it... -james ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Sat, Dec 30, 2000, 8:35 PM RE: Re: [ColorTheory] Color settings and upgrades on 12/30/00 4:13 PM, James Steincamp at steiny@cavtel.net wrote: > A perfect > device would require no such description as it would have a large enough > gamut to encompass the range of colors we are capable of perceiving and it > would precisely match our working RGB color space And throw in a display who's gamut is as large as the Working Space and output space. Plus we'd still need profiles. We still need to get to B (output space) which requires a description of A. We still need a B profile. > Managed Color has received a bum rap because a) in the beginning the > marketing folks allowed the hype to exceed what was delivered ['Color is > Child's Play' comes to mind or any product with 'EZ' in the name], and b) > many of those who believed the hype thought they were getting an automatic, > zero maintenance panacea for all their color problems. True indeed. >They failed to take > into account the fact that a CMS requires rigorous discipline to keep the > workflow under control (not at the service bureau or on press... where > process control is the norm), but in house where that had never been a > concern before. Well if not at the service bureau or press, where? All that is really needed beyond that is the display profile for each user. That's not a big deal to do accurately and at a low cost these days ($199). Input profiles from scanners would be nice but that's a service bureau issue as well as the end user who does his own scans. Assigning the display profile is not the most elegant way to pass the file into another space but it works. That's all we have when dealing with color negs anyway. > This is actually one feature that is particularly useful. It's nice to know > how someone else has chosen to process a file (usually, from my experience, > louse it up) on opening... without making a call (which could, at least in > my case, be half-way around the world in the middle of the night). The part of this that interests me is your comments about users usually lousing up something. Do these users screw up dealing with color settings in general, just color management settings or many uses within Photoshop that could potentially hose the file? Where does the responsibility to set up a system correctly so one doesn't hose a file lay? > But we will never know precisely where the imperfection lies as we have > neither a perfect device, nor a perfect instrument for measuring the > imperfections. We are left to muddle through as best we can. Measuring the "imperfections" far past what a human can discern isn't at all difficult to do with modest tools from several companies like Gretag or X-Rite. I'm not worried too much about the measurements although there are a number of issues to deal with (UV filters and such). > Or we deliberately distort the file in order to output as we would like. > Here we begin to move into the realm of subjectivity. Which do you like > better: Fujichrome's interpretation of color, or Kodachrome? I give that responsibility to the photographer. It's no different when I scan my color negs. I decide the rendering I want. But getting that onto print and screen as well (accurately) is the cool stuff that makes the job a LOT easier. > How many times has a client asked you to alter the color of some object in a > print because they 'feel' that the grass should be greener or that the faces > should be brighter. I'm the guy who used to take the pictures. I also did a lot of the imaging work back then when I was shooting. I don't recall this happening. But back then, we didn't have CMS and all the output was to a 4x5 or 8x10 chrome. Photoshop was just starting to work in CMYK! > Much of what 'color professionals' deal with on a > daily basis is pure trickery... and unfortunately lies beyond the pale of > what we can describe objectively. For that we have intuition, based on > previous (and yes, imperfect) knowledge. I'll put a good Photoshop photographer like Jeff Schewe go up after any color pro and he'll get what he wants a lot faster doing it himself with no special trickery other than Photoshop tools and some good profiles. I think every pro shooter should be shooting with a high end digital camera or scanning their own chromes or negs on a good scanner themselves or getting high bit raw scans from service providers. I wan them to do that work. It's their image. > Precisely what we have been doing without Color Management for years. Tinker > an learn. How many users of Photoshop who would like to deal with color can do this? > The problem that I have with Managed Color at this point is that with each > version of Photoshop and each incarnation of of ColorSynch and so on the > user is left grappling with (mostly undocumented) issues like the > differences between CMS engines and black point compensation as opposed to > learning about 'the thing in itself,' color. This is what Steve Jobs and the ColorSync product manager need to hear! What you are asking for should be supplied in some form by Apple. There is a ColorSync web page and there is some useful information there. I think Apple could have done more, and more quickly to update this fundamental information. There isn't a new incarnation of how Photoshop deals with color between versions 5 and 6. The basic idea is the same. The implementation is much better and the features are more powerful. > ColorSynch, etc. has come a > long ways in the last three years, but it will continue to change > dramatically (for the better) and require a fair amount of effort on behalf > of its users to keep up with the issues That's true for a LOT of stuff that runs on our computers! > Not really. If you remember in the early days of Managed Color there was the > idea of profiling the press itself... a scientifically sound notion. But > that fell by the wayside pretty quickly do to the cost involved... and now > we aim for the all-important MatchPrint (or whatever). That's one step > removed from the end product in the scenario that is most important to me. Maybe in this country. There are profile driven presses running splendidly in other countries of the world. Profiling to the Matchprint isn't all that bad either from my perspective. I don't own a Press or a Matchprint but I'd rather pay for one as a contractual agreement than have to be involved with the day to day operation of the presses. I'm just buying the stuff, not creating it so I'm on the other side of the fence. If I were the service provider I'd certainly be using a good profile for any RGB files I have to deal with. > There I am holding a contract proof saying... it looks like > the shadows still have to much 'insert your color here' without the benefit > of any scientific data, it's just me and what I understand about color. You still get to go to the press check. But if the stuff off the press and the contract proof (and your display and pre-proof to your Epson) all match within a reasonable degree, that's a nice way to get to that press check. > Your right that Mr. Margulis probably can't hit every 'print' on every > device the first time... but neither can you (or anyone else for that > matter). Bet I'll get it a lot quicker, more accurately and perhaps the first time, a lot of the time! > Naturally... or, to sharpern your point, as one SWOP press differs from > another. If they really were running SWOP (TR001), the shouldn't differ by very much at all. > One thing that I have learnt as an art director is that it can be > particularly dangerous to hand the client an ink-jet proof... it allows that > individual, with his or her grossly inadequate understanding of color, to > think that they are now qualified to start doing my job. But I can make my Epson look just like your Matchprint. What's wrong with a color comp that's accurate? Did anyone say anything about using the Epson AS a contract proof? If you don't agree to this and I want it, we don't have a contract. Both parties have to be in agreement. If you will not insure that my Epson can be used as a contract proof, that's fine with me. I'd want a final Matchprint anyway. But I can get my Epson to proof the color so I only need ONE Matchprint. I wouldn't show the client either or any proof unless it was under a daylight D50 box. But how many people don't do this? Andrew Rodney ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sat, Dec 30, 2000, 10:13 PM RE: [ColorTheory] confusing color management & color correction, was: Color settings and upgrades Long post warning. It has been a few months since we had a thrown down on color management vs. color correction, so here it is once again for those who have missed it in the past. James Steincamp writes: >Not really. If you remember in the early days of Managed Color there was the >idea of profiling the press itself... a scientifically sound notion. But >that fell by the wayside pretty quickly do to the cost involved... and now >we aim for the all-important MatchPrint (or whatever). There are reasons why press profiling never really took off. For one it wasn't necessary because of the ease of profiling contract proofing systems instead. And the vast majority of printers can't even get that right by providing their customers with accurate separation information (for either the press or the contract proofer). Ask them what dot gain for each channel, and printing inks information and they're like a deer in headlights. And keep in mind there are 32 some flavors of proofs that say MatchPrint on the back and the reason is because there are so many different kinds of press behavior. >It's key to understand that all the digital precision and profiling >information (and it's imperfections) ceases the moment the plates go on the >press and that means I'm still left waiting for that call in the middle of >the night to come down and do a press check... where I'm almost certain to >want to tweak. There I am holding a contract proof saying... it looks like >the shadows still have to much 'insert your color here' without the benefit >of any scientific data, it's just me and what I understand about color. Well wait a minute. Just in the previous paragraph you were explaining the glory of the MatchPrint. This is an argument about the inferiority of all proofs because we've needed things like press checks going back before a contemporary concept of "color management". >You're right that Mr. Margulis probably can't hit every 'print' on every >device the first time... but neither can you (or anyone else for that >matter). I don't think anyone would claim this, least of all Dan. His techniques are predominantly about fixing and enhancing images. Not about manual methods to short-circuit ICC profiles. If Dan is asked to do work that will ultimately print on a 12 color large format printer for both backlit media and matte coated media, he can still use the vast majority of his techniques, adapted in a way that doesn't make the same assumptions about this 12 color inkjet that he does about a four color press. I think he'll agree that given the expense of the media and the amount of time it takes for reprints, that accurate separation information is extremely helpful and he would prefer to use color management to assist in dealing with the pecularities of that device's output method than making educated guesses. This does not mean color management is used as a means to assist in fixing or enhancing his image. That is currently a uniquely human capability. But rather, it's used to ensure the desire of image reproduction occurs as close as possible in print. > A CMS system is dependent on having valid data about the process in >question; which means that you have to run a set of targets to provide the >proper data for the CMS. That means, in essence, that you have taken two >attempts to get to your finished result. That first print is distributed among each additional job. So this is a straw man argument. Unless you have a job that tests every aspect of color reproduction, high key, low key, a range of neutral gray, pastel, saturation, skin tones of different races, etc. then a manual process will take considerably longer. >I'm reasonably confident that a >competent numbers guy can hit any device (that behaves according to man's >present understanding of color) with two tries. I can do that pretty well >myself and I don't exactly have a decade's experience to guide me. Dan is a uniquely skilled and experienced individual and he has few equivalents in this industry. Dan's techniques are not designed to do what you suggest. They aren't about sizing up a device in two prints and being able to make the mental adjustment needed to know ink hues, ink/paper relationship, ideal black generation, ink limit, and dot gain for each channel. His techniques will produce better results without accurate separation information than if you don't use good technique, but I find it nearly impossible to believe that he would make any claim that he can size up the behavior of any randomly selected output device and get the exact results he wants just by making two attempts. >Naturally... or, to sharpern your point, as one SWOP press differs from >another. Well then it's not SWOP is it? SWOP is a rather specific thing since the days of TR001 are here. There is an expectation of colorimetric behavior (in Lab) for 938 CMYK values for SWOP, and in addition there is a visual comparison to a SWOP certified press sheet. The one that is off too much isn't SWOP. So the small variations I'll agree with, but those are minor and trivial to going from SWOP to a non-ink output process, or one that prints with 12-colors. > The process is exactly as yours... run a test. A numbers guy will >fall back on years of experience tweaking the numbers; a CMS guy will fall >back on the knowledge he has learnt about profiles and workflow. I totally disagree, but I'll go easy because you're not the first to fail to make a distinction between color correction and color management. It happens all the time. You are suggesting that a numbers person will be able to characterize any device in short order and make mental modifications and use that information to make appropriate separations. I think this is totally ludicrous and isn't supported by fact, unless you have weeks for the guy to make his trials and errors. Eventually yes he can become just as skilled on another form of output, but this implies that a good color and by the numbers person starts in the business and two hours on his first day he's a master color corrector. This kind of argument isn't even remotely compelling. Color correction and color management are not the same thing. They are not interchangable. They are non-competing techniques. One cannot be used in lieu of the other. If you suck at color correction, color management will only ensure the garbage on your monitor prints with the same visual horror in print. If every device in your workflow produces different results then the best color corrector will spend hours of unnecessary time using his skills to compensate for the pecularities of devices. Again they aren't the same thing. >Sure, the more you move away from a press and the CMYK we are all accustomed >to, the better the case for moving to a CMS. Yet most folks are still >focused on getting good color on press... where the bulk of money is >currently being made. I disagree. Most people in the graphic arts are trying to gear toward cross-media applications. That is, from the very beginning of a project, it's known it will be output to multiple forms of media. This is not the same thing as repurposing. There is a flawed concept that color management is the only thing that will make this possible and OK. There is also a flawed concept that refusal to adopt any form of color management will be OK. Both good color correction skills to fix the increasing number of messy imagery as well as color managment to contend with the every increasing diversity of output devices are needed. >[Aside] One thing that I have learnt as an art director is that it can be >particularly dangerous to hand the client an ink-jet proof... it allows that >individual, with his or her grossly inadequate understanding of color, to >think that they are now qualified to start doing my job. After all, 'I can >make a "quick change," right?' In all honesty there is a degree of finality >in showing the client a MatchPrint and the implicit understanding that more >revisions will cost $65 - $100 a pop. ;) Today's penalty is missed press times that cause a job to be pushed back instead of printed on schedule. Allow the customer to make whatever modifications they want. Each time they take the job away and bring it back means it will have to go through preflight again to make sure it will run through the workflow properly, there should be a charge for this service. It will have to be proofed again, and they will have to sign off on the proof. Matchprints don't work in the world of CTP where dot gain is at least 10% lower. Even Lo-Gain isn't low enough. People who move to CTP are dropping analog proofs like crazy and moving to digital proofs that can simulate a variety of output, including film based plate making as well as CTP, glossy coated paper to uncoated paper to newsprint. They are a lot more versatile and ultimately lead to fewer surprises. The client who enjoys making changes gets to pay in other ways. It's not necessary to scare them with a $100 proof. Scaring them with a three day delay in printing is far worse. >That's why the printers have resisted profiles... it puts the responsibility >for managing color in the lap of an often clueless end user... who in turn >will want to blame the printer when things go wrong. Printers resist EVERYTHING. They don't just resist profiles. They resist change period. If they resist profiles a little more it's just to be stubborn. Maybe the only legitimate reason for avoiding profiles it that it has a tendency to expose really bad process control. And in the world without trannies, who makes the separation anyway? Someone has to do it. Conventional workflows have a color house making the scans, and providing them as press ready CMYK files. How does this work when there is nothing to scan? How does it work when there is only RGB and there is no original to compare it to? The way to deal with this goes on both the customer and printer. Either the customer has to calibrate & profile their monitor, get correct separation information from the printer, and separate and color correct the image properly and take responsibility for it. Or the printer has to accept RGB images and separate them properly. Quite possibly the only thing that will make a printer freeze like a deer than asking for separation information is asking him to accept an RGB image. OK maybe asking him to guarantee it will print substantially the same (within the limits of the press of course) to what he saw on his calibrated and profiled monitor; and by the way the please respect the profile embedded in the image. I see the printer turning around and running away faster than he would from a dentist with a drill who has just run out of novocaine. And likewise there are a huge number of printer's customers who are in for a rude awakening also. They want to complain about color not matching their pesky $400 inkjet printers, and are full of blame for the printer, but for whatever crazy reason are unwilling to calibrate their monitor, profile their monitor, get accurate separation information for the press conditions in question, learn about the limitations of lithography, learn how to work with their printer, and learn decent color correction skills. So these printers and customers deserve each other and can keep on blaming each other, because that's exactly what they are asking for given the current trends. Chris Murphy ______________________________________________________________________ From: James Steincamp, INTERNET:steiny@cavtel.net Date: Sun, Dec 31, 2000, 1:52 AM RE: Re: [ColorTheory] Color settings and upgrades > And throw in a display who's gamut is as large as the Working Space and output > space. Plus we'd still need profiles. We still need to get to B (output space) > which requires a description of A. We still need a B profile. Not if things were indeed perfect. Hence the need for a strategy to cope with the problem. >> This is actually one feature that is particularly useful. It's nice to know >> how someone else has chosen to process a file (usually, from my experience, >> louse it up) on opening... without making a call (which could, at least in my >> case, be half-way around the world in the middle of the night). > > The part of this that interests me is your comments about users usually > lousing up something. Do these users screw up dealing with color settings in > general, just color management settings or many uses within Photoshop that > could potentially hose the file? Where does the responsibility to set up a > system correctly so one doesn't hose a file lay? That's the crux of the problem. In the past we just got randoms until we saw something we liked and then we ran with it. If the color didn't match I had an easy time pointing out the discrepancy and assigning blame. Now we have Photoshop which provides creative flexibility and the responsibility to understand what is happening to the file as we transform it. The closed loop is burst wide open... but the end user is still, more often than not, grappling with the basics. Photoshop and Illustrator provide plenty of opportunity for confusion (and little documentation)... each evolution of the apps presents changes that have significant impact on what comes out the other end of the workflow. What printer wants to accept that challenge (on tight margins)? Can you blame them, really? Take, for example, the new print dialogue in Photoshop. What's the average user supposed to think of the fact that, under 'Adobe Photoshop 6.0' you can specify a source space (document v. proof) and a print space (with profile and intent)... and yet under Color you have to option of setting the printer profile. It might seem simple to you... but where's the supplement to the manual that says, "so you want to manage color?" It's precisely those issues that give the user just enough rope to ensure that many of them will hang themselves. It's a lot more complicated than filling out a ballot... and look what headaches that created. >> Or we deliberately distort the file in order to output as we would like. Here >> we begin to move into the realm of subjectivity. Which do you like better: >> Fujichrome's interpretation of color, or Kodachrome? > > I give that responsibility to the photographer. It's no different when I scan > my color negs. I decide the rendering I want. But getting that onto print and > screen as well (accurately) is the cool stuff that makes the job a LOT easier. But as an art director I am responsible for making those calls. That's why you and I see the problem differently. I maintain that you cannot divorce a system (the CMS) from its utility. All I care about is the end result. >> Much of what 'color professionals' deal with on a daily basis is pure >> trickery... and unfortunately lies beyond the pale of what we can describe >> objectively. For that we have intuition, based on previous (and yes, >> imperfect) knowledge. >> > I'll put a good Photoshop photographer like Jeff Schewe go up after any color > pro and he'll get what he wants a lot faster doing it himself with no special > trickery other than Photoshop tools and some good profiles. I think every pro > shooter should be shooting with a high end digital camera or scanning their > own chromes or negs on a good scanner themselves or getting high bit raw scans > from service providers. I want them to do that work. It's their image. At least until the client pays for it. When I said trickery I didn't mean that color professionals were resorting to some black magic to get an exposure or proper image... but that color is subjective and influenced by psychological factors such as context and memory. Just last week I was preparing an add that had the color the client wanted (a blue-green, PMS 7473), but when viewed next to an area of yellow it 'appeared' wrong... shifting in color. The client swore that I had the color wrong. It wasn't colormetric data that convinced him... but taking the PMS swatchbook and placing it next to his logo (his contract proof, or his anchor on reality, so to speak) and then taking the two and placing them next to the MatchPrint. He was stunned that his eyes had deceived him. One of my profs at RIT recalled a studio shoot where the client swore his beer was more amber than it really was. This stuff is part and parcel of what color professionals deal with. We need more than just a CMS, but the tools to understand. >> Precisely what we have been doing without Color Management for years. Tinker >> an learn. > > How many users of Photoshop who would like to deal with color can do this? And how many PS users can you hand a copy of ProfileMaker and a SpectroLino and expect them to navigate the issues of Managed Color without a fair amount of hand-holding? Then bear in mind that a growing trend in graphic design is for the poor office secretary to get a copy of PageMaker, a color printer and a 'promotion' to graphic designer. > But I can make my Epson look just like your Matchprint. What's wrong with a > color comp that's accurate? We are looking at the problem slightly differently... but we're more in agreement than not. I think if you go back to my original post and replace the term 'describe imperfection' with 'spectral characteristics' and 'corrects' with neutralise... I think you will have to agree that I got it right. I merely ascribed to ColorSync the objective of establishing a Managed Color workflow. > Then there is profiling... which in a Managed Color environment typically > means to create a ColorSync (or ICC) profile. Profiling is an acknowledgement > that no device is perfect (even when calibrated) and serves as an attempt to > describe a device's (your monitor in this case) imperfections. ColorSynch then > attempts to eliminate those imperfections from the process so that you are not > changing your wonderful RGB file based upon incorrect assumptions. For example > you may try to color correct in a room that has fluorescent lights... you're > seeing a lot of green in that case... and correcting under those conditions > would likely leave your file too magenta under tungsten lighting. A profile > attempts to correct that situation. But I still maintain that: 1. Without allowing the user to 'correct' the inherent 'imperfections' of the devices in a workflow... a CMS has no utility and no value. 2. A Color Managed Workflow is also only as strong as its weakest link... be that bad process control, inaccurate profiles or a user that is ignorant of the goals and methods of the CMS. 3. Aspiring color professionals and dedicated amateurs are better off learning the basics of color, image correction, etc. before making an investment in CMS tools. 4. CMS should not be presented as simple. It's not a 'point and shoot' proposition. It's a form of process control and it requires effort to implement and perpetuate. You can't have half a CMS and expect it to work. 5. As the workflow diversifies to include alternatives to traditional press output the value of a CMS increases... and so does the knowledge and energy required to ensure that the results are indeed the desired results. -james ______________________________________________________________________ From: James Steincamp, INTERNET:steiny@cavtel.net Date: Sun, Dec 31, 2000, 1:52 AM RE: Re: [ColorTheory] confusing color management & color correction, was: Color settings and upgrades > Long post warning. It has been a few months since we had a thrown down on > color management vs. color correction, so here it is once again for those > who have missed it in the past. eh? Did you actually bother to read my posts? > There are reasons why press profiling never really took off. For one it > wasn't necessary because of the ease of profiling contract proofing > systems instead. I'm afraid that it had a lot more to do with cost and friction from the press owner than the relative ease of profiling the MatchPrint. > Well wait a minute. Just in the previous paragraph you were explaining > the glory of the MatchPrint. This is an argument about the inferiority of > all proofs because we've needed things like press checks going back > before a contemporary concept of "color management". I wasn't holding up the MatchPrint as anything particularly wonderful... just as a necessary element in the way that most printers go about conducting business. Yet whether I have reached that point via a CMS or the 'old fashioned' method... there I am... comparing it to what rolls off the press. I'm eyeballing it. >> Your right that Mr. Margulis probably can't hit every 'print' on every >> device the first time... but neither can you (or anyone else for that >> matter). > I don't think anyone would claim this, least of all Dan. His techniques > are predominately about fixing and enhancing images. Not about manual > methods to short-circuit ICC profiles. If Dan is asked to do work that > will ultimately print on a 12 color large format printer for both backlit > media and matte coated media, he can still use the vast majority of his > techniques, adapted in a way that doesn't make the same assumptions about > this 12 color inkjet that he does about a four color press. I've already said that, as you move away from a traditional prepress environment, the case for CMS becomes more compelling. >> A CMS system is dependent on having valid data about the process in >> question; which means that you have to run a set of targets to provide the >> proper data for the CMS. That means, in essence, that you have taken two >> attempts to get to your finished result. > > That first print is distributed among each additional job. So this is a > straw man argument. Not if your printer hasn't run the targets previously... and I've yet to met a major printer that's prepared to say... 'yeah, here you go, it's all on disk,' for either their in house proofing system or for their presses. That's your cost (in time and money) now. >> Naturally... or, to sharpern your point, as one SWOP press differs from >> another. > > Well then it's not SWOP is it? SWOP is a rather specific thing since the > days of TR001 are here. There is an expectation of colorimetric behavior > (in Lab) for 938 CMYK values for SWOP, and in addition there is a visual > comparison to a SWOP certified press sheet. The one that is off too much > isn't SWOP. So the small variations I'll agree with, but those are minor > and trivial to going from SWOP to a non-ink output process, or one that > prints with 12-colors. Yeah... that sounds great. And how many printers actually meet those specs? Every press and press operator is different. Being ISO or DIN compliant doesn't make a company great, either. >> The process is exactly as yours... run a test. A numbers guy will >> fall back on years of experience tweaking the numbers; a CMS guy will fall >> back on the knowledge he has learnt about profiles and workflow. > > I totally disagree, but I'll go easy because you're not the first to fail > to make a distinction between color correction and color management. I believe you are the one who failed to understand the point that I was making. I do understand the difference between color correction and color management. The point of my exchange with Andrew is that you cannot divorce a Color Management System from its objectives, it's utility. The objective of a CMS is to eliminate the negative impact that various devices, with their varying behavior, have upon an image as it moves through the workflow. Maintaining a CMS, is in turn about managing the process so that it remains stable. There's a lot of work there. The origin of our dialogue was his objection to my use of the term 'imperfection.' I had said that a profile attempts to describes a device's imperfection... as all devices are imperfect. > It happens all the time. You are suggesting that a numbers person will be > able to characterize any device in short order and make mental modifications > and use that information to make appropriate separations. Not as precisely as a profile. But I'll put my money... any day... on an experienced numbers guy when matched up with a total newbie and his CMS. Spending US$7.5k on the Gretag toys (nice stuff, though) does not mean consistent output. It takes effort to set up a workflow and manage it... and it takes knowledge of CMYK and the prepress industry. I'm tired of seeing Profiling/CMS presented as 'simple' solution with the same reckless abandon that owning a Minolta Maxxum takes you to the same level as a professional photographer. There's something to be said about understanding the limitations of the system and having the knowledge to bridge the gaps. > I disagree. Most people in the graphic arts are trying to gear toward > cross-media applications. That is, from the very beginning of a project, > it's known it will be output to multiple forms of media. This is not the > same thing as repurposing. It's moving that way. Slowly. > There is a flawed concept that color management is the only thing that > will make this possible and OK. There is also a flawed concept that > refusal to adopt any form of color management will be OK. Both good color > correction skills to fix the increasing number of messy imagery as well > as color managment to contend with the every increasing diversity of > output devices are needed. Hence the necessity of mastering the material in Mr. Margulis' book. A CMS does not excuse one from learning about color. > They [printers] don't just resist profiles. They resist change period. Now there's one thing we can agree on. But why are printers so conservative? It's because the gross margins have fallen significantly in the last decade and yet the price of the equipment continues to rise. The good printers have adopted some form of Total Quality Management... to their credit. The rest have been bought up, gone belly up or have started to look for greener pasture elsewhere. Look at what's been thrown their way in just the last few years: PDF, PostScript 3, CTP, DI, HiFi printing... and I'm just getting started. The guys that have to deal with this stuff don't get paid mega-bucks like the Perl-loving, GNUmatic Linux geek your bank hires to run and protect its network. But the complexity involved isn't much different. > The way to deal with this goes on both the customer and printer. Either > the customer has to calibrate & profile their monitor, get correct > separation information from the printer, and separate and color correct > the image properly and take responsibility for it. Or the printer has to > accept RGB images and separate them properly. Yet, if they don't have the spectral information about the source of the RGB image... then what? A good CMS is about controlling and quantifying the variables; as the information grows sparse the system begins to unravel. In this case you're asking the printer to rely on the client, who certainly has the financial motive to point the finger at the printer when things go wrong. > So these printers and customers deserve each other and can keep on > blaming each other, because that's exactly what they are asking for given > the current trends. That's the way it's been for a long time. I didn't say that it was pretty or desirable. ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sun, Dec 31, 2000, 6:18 AM RE: Re: [ColorTheory] confusing color management & color correction, was: Color settings and upgrades >I'm afraid that it had a lot more to do with cost and friction from the >press owner than the relative ease of profiling the MatchPrint. Cost doesn't have a lot to do with it. If it did, they would profile their contract proofing system, which by your admission is easier and cheaper. Yet the vast majority of printers don't do this. >I've already said that, as you move away from a traditional prepress >environment, the case for CMS becomes more compelling. There are no traditional prepress shops anymore. Find me a service bureau that makes it's living only on making scans of film and making film to make plates. They are doing a lot more than this these days in order to survive. They are being asked to prepare files for multiple kinds of output. They do retouching. They produce digital proofs in addition to analog proofs. The largest growing segment in digital printing is at prepress houses - they are buying the high-end color copiers and even digital presses. Second, color management happened even in those traditional prepress environments. It happened in the form of algorithms in scanners that converted RGB data into CMYK on the fly. It happened in the form of intimate understanding of specific kinds of printing on the part of the scanner operator. It happened with analog proofs as a built-in function. >Not if your printer hasn't run the targets previously... and I've yet to met >a major printer that's prepared to say... 'yeah, here you go, it's all on >disk,' for either their in house proofing system or for their presses. Exactly my point. Shame on them. To want accurate separation information for either their press or their contract proofing system and for them to be totally unable to provide it is TERRIBLE. >> Well then it's not SWOP is it? SWOP is a rather specific thing since the >> days of TR001 are here. > >Yeah... that sounds great. And how many printers actually meet those specs? Yeah and again that's my point. If you follow SWOP, you will be able to match the SWOP Certified Press Sheet and TR001. It's been repeated, by SWOP Inc., on different presses over several years time. So it's not impossible. >Every press and press operator is different. Being ISO or DIN compliant >doesn't make a company great, either. When it comes to advertising it makes a big difference. Hence organizations (which are effectively movements) like SWOP, dMACS, and DDAP. For a company to claim they print per SWOP, yet don't have TR001 behavior, and can't match a SWOP Certified Press Sheet is making a false claim. A company making a false claim isn't a great company anyway. > The point of my exchange with Andrew is that you cannot divorce >a Color Management System from its objectives, it's utility. The objective >of a CMS is to eliminate the negative impact that various devices, with >their varying behavior, have upon an image as it moves through the workflow. >Maintaining a CMS, is in turn about managing the process so that it remains >stable. There's a lot of work there. CMS isn't just about that. It's about even the simplist thing even when there is only one form of output - the press. You still need a means to make a separation. If that separation isn't going to happen with a tranny on a drum scanner, then you must have some kind of accurate separation information, or you need extra time to manually color correct to FIX an image that was improperly separated. Color management can be simple or complex. It's not inherently simple or complex. It depends on what you do now. If your current workflow is simple, color management can be simple. >The origin of our dialogue was his objection to my use of the term >'imperfection.' I had said that a profile attempts to describes a device's >imperfection... as all devices are imperfect. Compared to what? To refer to them as imperfect means there is something perfect in order to have basis of comparison that lets you conclude perfection or imperfection. I wouldn't call them imperfect. I would call them different, as in unique. >> It happens all the time. You are suggesting that a numbers person will be >> able to characterize any device in short order and make mental modifications >> and use that information to make appropriate separations. >Not as precisely as a profile. But I'll put my money... any day... on an >experienced numbers guy when matched up with a total newbie and his CMS. OK now this is going down a totally different road than you were originally now that you are comparing an experienced person to an inexperienced person (as if that's even a fair comparison); and you're making the same mistake I'm complaining about at the beginning of this post. You are comparing an experienced numbers person (color correction) with color management. They are not comparable in this fashion. They are two different skills and have two different purposes. The previous argument left it pretty much going down the road of saying a skilled by the numbers person would need only two attempts to get color right on a random output device they previously had no experience with. That's a different argument that the above. >Spending US$7.5k on the Gretag toys (nice stuff, though) does not mean >consistent output. It takes effort to set up a workflow and manage it... and >it takes knowledge of CMYK and the prepress industry. Again, it depends on the situation. If you're a graphic designer, then you really only need to know two things about color management: 1.) The procedure for calibrating and profiling the monitor. 2.) The settings in Photoshop that you should be using, most important of which is getting the most accurate CMYK information plugged in. This would take at most couple of hours of training for someone who is already suitably experienced with offset printing. It's not that big of a deal. And not that big of an expense. All you'd have to do is save $500 in extra MatchPrints for example. Something that could be done in a few days to a week for the reasonably occupied designer. >I'm tired of seeing Profiling/CMS presented as 'simple' solution with the >same reckless abandon that owning a Minolta Maxxum takes you to the same >level as a professional photographer. It can be that simple. It only gets more complicated if you want to profile scanners, digital cameras, and then output devices. Inkjet printers aren't that difficult for the most part. Printing presses are much more complicated to profile, but there's not reason for the average person to do that anyway. Your PRINTER should be able to give you suitable separation information and the vast majority can't even do that. >> I disagree. Most people in the graphic arts are trying to gear toward >> cross-media applications. That is, from the very beginning of a project, >> it's known it will be output to multiple forms of media. This is not the >> same thing as repurposing. > >It's moving that way. Slowly. Where have you been? It's the talk of the industry, in nearly every printing/pre-press magazine, and trade show or conference. Cross-media application demand by ad agencies is rapidly increasing because clearly they want to target multiple demographics which means using multiple forms of media. And they want to do it all at the same time for maximum impact. >Hence the necessity of mastering the material in Mr. Margulis' book. A CMS >does not excuse one from learning about color. I haven't heard anyone imply otherwise. >> They [printers] don't just resist profiles. They resist change period. > >Now there's one thing we can agree on. But why are printers so conservative? >It's because the gross margins have fallen significantly in the last decade >and yet the price of the equipment continues to rise. This is a valid excuse for being conservative on equipment purchases. But it's not an excuse for the inability to provide reasonable separation information. We're talking about 14 measurements (minimum, up to 61) and using Photoshop to build the "profile". Only 9 measurements are potentially unique for a print shop that doesn't have a colorimeter or spectrophotometer. The other measurements only need a densitometer. Something every printer should have by now (you'd think). I'm sure they can find someone who will get them their 9 xyY measurements for ink hues if they really wanted to. >> The way to deal with this goes on both the customer and printer. Either >> the customer has to calibrate & profile their monitor, get correct >> separation information from the printer, and separate and color correct >> the image properly and take responsibility for it. Or the printer has to >> accept RGB images and separate them properly. > >Yet, if they don't have the spectral information about the source of the RGB >image... then what? Colorimetric information. There are only a couple of systems that use spectral information at this point. If the customer doesn't embed a profile in their image, *and* doesn't calibrate and profile their display so that others know how they were seeing the image, then the customer isn't in a position to complain. They can pay for extra MatchPrints and either sign them or not. But complaining about how a proof looks compared to their monitor should draw no sympathy. If they are serious about solving the problem, they will get serious about calibration/profiling, embedding and using a printer that knows how to deal with images with embedded profiles. > A good CMS is about controlling and quantifying the >variables; as the information grows sparse the system begins to unravel. If the image doesn't have a profile in it, the printer can still open the image on their wonderfully calibrated/profiled monitor, color correct as needed in RGB, make an appropriate separation to CMYK, and generate the proof. There is no unraveling and it's not a big deal. There is no way to not use color management anyway. If there is no profile embedded in the image, one gets assumed by all versions of Photoshop since version 2.5. Something has to be defined for RGB before a conversion to CMYK can occur. You can't just choose nothing for RGB and get a separation. > In >this case you're asking the printer to rely on the client, who certainly has >the financial motive to point the finger at the printer when things go >wrong. They each of sufficient financial motive to blame each other. If each does their minimum duty, errors and complaints regarding color are far fewer, and when they do come up, troubleshooting how they occured (and who's to blame if you want to look at it that way) is easier because responsibilities have been predefined. Right now we're working with mahem which is why it's even possible to point fingers at each other in the first place. Customer responsibilities regarding CMS (prerequisites are knowledge of offset printing and its limits, and good color correction skills such as what Dan teaches): 1.) Calibrate and profile monitor. Cost $199 for very high quality capability, and maybe an hour to install and familiarize yourself with on-screen instructions and go through the first event. Subsequent events are maybe 10 minutes. 2.) Embed correct profiles in your images (Photoshop 6 makes this nearly fool proof) 3.) Ensure accurate separation information supplied by the printer has been plugged into Photoshop/Illustrator before doing the conversion. This ensures CMYK previews are accurate as well as any RGB-CMYK conversions. Printer responsibilities: 1.) Good process control. 2.) Ability to provide customers with press/contract proof separation information. There really isn't even a need for the printer to make separations unless they want to, or their customer wants them to. In that case they need to respect embedded profiles, plug in accurate separation info into Photoshop, and then convert the image. Naturally either process will still benefit from process specific tweaks to the resulting CMYK image. Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sun, Dec 31, 2000, 6:37 AM RE: Re: [ColorTheory] Color settings and upgrades >Take, for example, the new print dialogue in Photoshop. What's the average >user supposed to think of the fact that, under 'Adobe Photoshop 6.0' you can >specify a source space (document v. proof) and a print space (with profile >and intent)... and yet under Color you have to option of setting the printer >profile. This isn't new with Photoshop 6. It's been a problem ever since Adobe's PSPrinter version 8 I believe (which they send to Apple and Apple makes Laserwriter 8 which has the same problem). The Color pop-up is designed for applications that don't support color management, so you do it in the driver instead. For Photoshop you ignore this section of the driver. But this is a huge huge huge complaint that the color management users and geeks have had with both Adobe and Apple as long as I can remember. Maybe I'm just more vocal about giving them royal crap for it. But I think they deserve even more for the total lack of driver documentation. It took conversations with current and former ColorSync engineers and Adobe programmers as well to finally understand what the deal was. It's terrible user interface. > I maintain that you cannot divorce a >system (the CMS) from its utility. I don't know what you mean by this. >Just last week I was preparing an add that had the color the client wanted >(a blue-green, PMS 7473), but when viewed next to an area of yellow it >'appeared' wrong... shifting in color. The client swore that I had the color >wrong. It wasn't colormetric data that convinced him... but taking the PMS >swatchbook and placing it next to his logo (his contract proof, or his >anchor on reality, so to speak) and then taking the two and placing them >next to the MatchPrint. He was stunned that his eyes had deceived him. This shows his lack of experience with color. Color would be easier if more people understood how easy it is to fool human color perception. >One of my profs at RIT recalled a studio shoot where the client swore his >beer was more amber than it really was. This stuff is part and parcel of >what color professionals deal with. We need more than just a CMS, but the >tools to understand. Well they don't get it in art school apparently. CMS isn't even a factor in what you're talking about because right now there isn't a system that compensates for the effect you're referring to above (simultaneous contrast). There is a method being worked on that can predict this, but at this rate I'll be retired by the time we see this. (And if you knew how old I was you'd know that meant a really long time.) >And how many PS users can you hand a copy of ProfileMaker and a SpectroLino >and expect them to navigate the issues of Managed Color without a fair >amount of hand-holding? Then bear in mind that a growing trend in graphic >design is for the poor office secretary to get a copy of PageMaker, a color >printer and a 'promotion' to graphic designer. Well I sympathize but they get what they get. It's just like if I'm going to try and do surgery on myself instead of seeing a doctor. If I'm going to try and save a buck and cut myself open, then I pretty much deserve what I'm gonna get by cutting myself open. >3. Aspiring color professionals and dedicated amateurs are better off >learning the basics of color, image correction, etc. before making an >investment in CMS tools. I don't think it's an issue of one over the other or one before another. I think both are important. In *this* context I agree that these things are more important and should come first before color management. However, for that secretary using Pagemaker, he's in a pretty decent position to take advantage of the more automated "consumer" versions of color management to get sort of what he sees on his monitor on his inkjet printer and not needing to know anything about color basics. Chris Murphy ------------------------------------------------------------------------------- From: "richardkenward", INTERNET:richardkenward@onetel.net.uk Date: Sun, Dec 31, 2000, 7:19 AM RE: Re: [ColorTheory] confusing color management & color correction, was: Color settings and upgrades Well for my money, Chris and Andrew have put the argument extremely well and very convincingly...as usual! What I find quite extraordinary is how the print industry in general, manage to find such great difficulty in creating the environment for predictable, reliable, efficient working. If they got their act together, then there is no longer an excuse for the file providers to send in shoddy work. That should stop the 'buck ' passing and allow serious professionals to get down to producing excellent work. Regards Richard Richard Kenward Richard Kenward Digital Imaging +44 (0)1873 890670 ------------------------------------------------------------------------------- From: "Dave Badger", INTERNET:dbadge@worldnet.att.net Date: Sun, Dec 31, 2000, 12:14 PM RE: Re: [ColorTheory] confusing color management & color correction, was: Color settings and upgrades > Printing presses are > much more complicated to profile, but there's not reason for the average > person to do that anyway. Your PRINTER should be able to give you > suitable separation information and the vast majority can't even do that. I'm one of the vast majority who can't provide accurate separation information for our new digital press. We hired one of the best pressman in town and profiled the press, but it was useless. When I talk with the pressman he says he can't match Iris proofs and gives me an endless list variables as to why presses can't be profiled; ink tack, ink viscosity, blanket pressure and age, paper color and stock, layout of the particular piece, and on and on. He has all kinds of tricks for making the job look good and sees printing more of an art then science, saying prepress people (and art directors, designers, etc) don't understand what happens when its time to put ink on paper. Having always lived in the theoretical world of prepress and not knowing the press, I'm starting to believe him. > This is a valid excuse for being conservative on equipment purchases. But > it's not an excuse for the inability to provide reasonable separation > information. We're talking about 14 measurements (minimum, up to 61) and > using Photoshop to build the "profile". Only 9 measurements are > potentially unique for a print shop that doesn't have a colorimeter or > spectrophotometer. The other measurements only need a densitometer. > Something every printer should have by now (you'd think). I'm sure they > can find someone who will get them their 9 xyY measurements for ink hues > if they really wanted to. So what good are these measurements if the press is running all different stocks of paper, and press conditions change as a result of all the variables listed above. Its easy to say the pressman just doesn't want to adhere to process controls and standards, but hard to argue with quality results and satisfied customers he produces. So I could hand out Photoshop sep profiles, but he's always going to have to "fix" it on the press anyway. Dave Badger ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Sun, Dec 31, 2000, 1:06 PM RE: Re: [ColorTheory] Color settings and upgrades on 12/30/00 11:51 PM, James Steincamp at steiny@cavtel.net wrote: >> And throw in a display who's gamut is as large as the Working Space and >> output >> space. Plus we'd still need profiles. We still need to get to B (output >> space) >> which requires a description of A. We still need a B profile. > > Not if things were indeed perfect. Hence the need for a strategy to cope > with the problem. I think we would still want a really wide gamut display so we could see what we are doing. This allows the color correction verses CMS to take hold as a user could see what he's doing as he corrects (along with the info palette). The display when properly calibrated is just another tool to do the color correction work. But when the output or working space has a gamut that is larger than our displays, this can in rare cases get dicey. Adobe put a band aid in Photoshop 6 to allow our preview to desaturate in order to allow us to see the effect of out of gamut colors altering as we edit them. Downside is the soft proof goes out the window. What we need are wider gamut display technology. I don't see that happening in the near future. This of course is less an issue for the 4 color CMYK group then the very wide gamut RGB output devices. But there are still some nice very saturated cyan's in a CMYK space that can't be displayed on an RGB monitor. > That's the crux of the problem. In the past we just got randoms until we saw > something we liked and then we ran with it. If the color didn't match I had > an easy time pointing out the discrepancy and assigning blame. I can't argue that this wouldn't work and eventually produce fine output. My worry is the time, energy and money to get to this final point. My experience is that with a good set of profiles and a reliable output device, I can get to my final with one, perhaps two proofs max. So getting to the final goal with no more than a 2nd proof (and usually hitting it the first time), I'm pretty happy. > Now we have > Photoshop which provides creative flexibility and the responsibility to > understand what is happening to the file as we transform it. The closed loop > is burst wide open... but the end user is still, more often than not, > grappling with the basics. Photoshop and Illustrator provide plenty of > opportunity for confusion (and little documentation)... each evolution of > the apps presents changes that have significant impact on what comes out the > other end of the workflow. Yes. Photoshop is just over 10 years old. It's been doing CMYK for at least 8. We have an open, flexible system that anyone with $600 to purchase Photoshop can use. There is a downside to providing such power to the novice. I think that any such tool in the wrong hands (your SLR for example) will be a recipe for potential problems. I'm not sure that having open, powerful systems available to any user is necessarily a bad thing. It boils down to the service provider again. We can't put the toothpaste back in the tube. As a service provider, it's not possible to ignore all the Photoshop users (both advanced and novice). So we either have to get with the program and get these lesser sophisticated users working properly OR ignore them and their business OR provide some training and hand holding. Ignoring how Photoshop 5 and 6 deals with color isn't a very viable solution. It's not going away. But the good news is that the stuff works! Yes, it's complicated. Yes, it's going to take training. > What printer wants to accept that challenge (on tight margins)? Can you > blame them, really? Not at all. I only blame the printers who decide that a viable solution which has existed for the last 3 years doesn't work based on an article they read or just because it involves changing a single setting somewhere in their shop. I've actually seen the latter mindset many times. That's another story however... > Take, for example, the new print dialogue in Photoshop. What's the average > user supposed to think of the fact that, under 'Adobe Photoshop 6.0' you can > specify a source space (document v. proof) and a print space (with profile > and intent)... and yet under Color you have to option of setting the printer > profile. It might seem simple to you... It wasn't simple. I was able to figure out a lot of it due to experience. But I did read the manual and some on line articles. I do have the ear of the Adobe team as I'm a Photoshop beta and can ask them what it's all about in an attempt to create PDF's and explain to others what's going on. I absolutely agree that the one area in the Photoshop team that is very weak is the documentation group. They did better in Photoshop 6 than 5 (which isn't saying a lot considering how little information there was about Photoshop 5 when it first shipped). The stuff does take some effort to learn. So does the Channel Mixer, the image size dialog, the Extract Command. It's just that when the topic is something to do with color, so many people throw up their hands and just say "too complicated, let's forget it." Hell, I'll take the CMS features in Photoshop 6 over the Extract and even the Channel Mixer any day. > but where's the supplement to the > manual that says, "so you want to manage color?" I purchased Photoshop 1.07 in April or May of 1990. I read the entire manual (there were no books, web pages or anything else back then). I don't have the original manual but I'll bet it's as big as the Photoshop 6 manual. Back then, you HAD to read about (and Adobe had to explain) exactly what the Clone Tool did. They did a pretty good job. I have no idea when Adobe started to slack off on writing good manuals. Maybe they wanted the 3rd party authors to have a field day. The info you seek is out there but NOT in the manual! > But as an art director I am responsible for making those calls. I've worked with a lot of AD's and a lot of other photographers. Some don't have a clue about color. Others do. You don't tell Herb Ritz how the color should look in his images. There should however be a creative exchange between the two. As a Photographer, I'd never think of telling an AD my impressions of what type face the final ad should be in. But the AD knew I knew about lighting and the control over my image. Every job and relationship is different of course. However, when it comes to final color, IF I had to pick one individual to control this (assuming this person knew what they were doing), I'd have to pick the photographer. > When I said trickery I didn't mean that color professionals were resorting > to some black magic to get an exposure or proper image... but that color is > subjective and influenced by psychological factors such as context and > memory. Exactly and I like to rely on the photographers memory. And when dealing with color negs (something anyone in prepress would fall apart dealing with) the only option is a calibrated display and someone to set the color rendering of the scene. There's no reference (a chrome) to allow someone to match. And why would I want to match the chrome? Sometimes I don't. What about all the fine images shot on digital cameras? What do we match and who is in charge of deciding what is the right match? Again, I'd go with the Photographer assuming they knew what they were doing. > He was stunned that his eyes had deceived him. Not all that uncommon! > One of my profs at RIT recalled a studio shoot where the client swore his > beer was more amber than it really was. This stuff is part and parcel of > what color professionals deal with. We need more than just a CMS, but the > tools to understand. Having the ability to measure makes the ambiguity a bit easier to deal with (by removing that ambiguity). That's not to say a Spectrophotometer is a tool that will always solve such problems. But it's a tool that can get you to the final quicker in some cases. It's also good ammo for use on that client! > And how many PS users can you hand a copy of ProfileMaker and a SpectroLino > and expect them to navigate the issues of Managed Color without a fair > amount of hand-holding? How many printers (pressman) can you hand a copy of Photoshop to and have them do dust busting. How many Art Directors can you had a 4x5 camera and a cereal box to and get an exposed 4x5 with the image square using swings and tilts? Forgive me but it's kind of a stupid question. No one is saying that you can hand a Spectrolino to anyone with no training and end up with anything but frustration. Why would anyone expect otherwise? In about an hour I can show someone how to use the Spectrolino to build a really good profile in ProfileMaker Pro. That person could probably do the same with a day of manual reading and playing. What I keep failing to understand is why all the CMS nay-sayers raise this issue you have. It's as if a new technology that works also has to be so intuitive that anyone that can tie their show can automatically be able to work with the product. That wasn't the case the first time I saw Photoshop's Curves command yet I didn't ignore it, curse at Adobe for providing that tool and just keep using "Contrast and Brightness" because it seemed so much simpler to use. I sat down and learned what curves do and never looked back at the work involved. > 1. Without allowing the user to 'correct' the inherent 'imperfections' of > the devices in a workflow... a CMS has no utility and no value. Agreed. Chris has talked about color correction and CMS and how the two are not the same. A CMS will insure that a really ugly file prints really ugly. > 2. A Color Managed Workflow is also only as strong as its weakest link... be > that bad process control, inaccurate profiles or a user that is ignorant of > the goals and methods of the CMS. True of any process. As an AD, what would be your reaction if you had your photographer hold one sheet of 4x5 and run the first normal E6. After seeing that normal image, you asked the lab to run that film + a half a stop based on that first 4x5. But what if the film came back a third stop dark because the process control of the E6 was flaky? You'd have kittens. With really, really bad amateur labs, that scenario is possible I think. With any pro lab, that wouldn't happen unless something really failed (and that happens VERY infrequently with pro labs or they don't stay in business very long). So why is it that when it comes time to print 100,000 copies of that one 4x5, process control isn't all that critical? Why do the printers get off the hook? Bottom line, any process needs process control or we just have chaos. > 3. Aspiring color professionals and dedicated amateurs are better off > learning the basics of color, image correction, etc. before making an > investment in CMS tools. No, it should happen at the same time! Learning about a CMS allows a user to learn a LOT about color and reproduction. It's silly teaching about numbers or correcting when you don't have a good handle on the output process (which profiles allow us to have) or a calibrated display. > 4. CMS should not be presented as simple. It's not a 'point and shoot' > proposition. It's a form of process control and it requires effort to > implement and perpetuate. You can't have half a CMS and expect it to work. It's not simple to implement (build profiles and so forth) but once in place it's a LOT simpler then with no profiles or properly configured copy of Photoshop. Also, it's not about process control. That's the job of the people doing the output. As long as I have a profile that reflects your proofer, and that proofer remains consistent, I could care less what you do to linearize or calibrate the device daily or how you replenish your chemistry or the plots of your control strips. Yes, a CMS expects process control. As I said, I expect process control even when there isn't a CMS in sight (like at my E6 lab). Process control is immaterial here, it's expected. With or without a CMS, a process that isn't in control is a recipe for disaster. Andrew Rodney ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Sun, Dec 31, 2000, 1:32 PM RE: Re: [ColorTheory] confusing color management & color correction, was: Color settings and upgrades on 12/31/00 5:17 AM, richardkenward at richardkenward@onetel.net.uk wrote: > What I find quite extraordinary is how the print industry in general, > manage to find such great difficulty in creating the environment for > predictable, reliable, efficient working. It's pretty amazing. We can't and should NOT blame the entire industry. By and large, the majority fit into your description but there are more and more shops implementing really solid process control and using profiles. What's really interesting is I'm seeing this happening quickly with labs dealing with RGB processes (people with Lambda and equipment that is as costly as much of what you'll see in conventional print shops). I think these labs have grown from the traditional photo lab so their midset (and openness for change) is in a different place than the traditional print shops. It's also interesting to see many print and prepess operations opening up digital studios. Now with their scanning backs or other such systems, they get to deal with RGB. I spent a few days at such an operation in Texas this year. Profiling the digital camera, the Epson used for "pre-Proofs" and building a CMYK profile for their contract proofing system wasn't much work (a day). What was interesting was when I handed the prepress department an RGB file from the camera (with input profile) but also gave them a CMYK file from the same image I converted using the custom profile. The two printed side by side and there was no surprise on my part that the file the Prepress department separated (using Photoshop CMYK defaults; no surprise here) looked pretty bad next to my custom conversion. The owner of the shop agreed that my sep was significantly better and all RGB to CMYK conversions from the studio would now be conducted using this method. When the plant manager asked what would be required for PrePress to do the same on their machines, I told them all they had to do was load my CMYK ICC profile into their copies of Photoshop. He refused saying it was "too much work" and what was really frightening was that the owner agreed! The output was staring them in the face but the prospect of changing a single setting in Photoshop (from what is just a default and not a very good one) was just too scary for them to do. To this day, the digital studio is fully color managed (and doing more and more work I should ad) and prepress goes along in ignorant bliss while their workflow for capture decreases. Andrew Rodney ------------------------------------------------------------------------------- From: "richardkenward", INTERNET:richardkenward@onetel.net.uk Date: Sun, Dec 31, 2000, 2:03 PM RE: Re: [ColorTheory] confusing color management & color correction,was: Color settings and upgrades > on 12/31/00 5:17 AM, richardkenward at richardkenward@onetel.net.uk wrote: > > > What I find quite extraordinary is how the print industry in general, > > manage to find such great difficulty in creating the environment for > > predictable, reliable, efficient working. I should have said 'generally' for that is what I actually meant, there are thankfully always exceptions ! > It's pretty amazing. We can't and should NOT blame the entire industry. By > and large, the majority fit into your description but there are more and > more shops implementing really solid process control and using profiles. > What's really interesting is I'm seeing this happening quickly with labs > dealing with RGB processes (people with Lambda and equipment that is as > costly as much of what you'll see in conventional print shops). I think > these labs have grown from the traditional photo lab so their midset (and > openness for change) is in a different place than the traditional print > shops. Over here in the UK I have yet to come across a Photo Lab that uses anything better than sRGB for running their digital printing gear, be it Kodak LED or Lambda etc. Now this seems strange that with the gamut available from traditional (modified for the short laser exposures) RA4 colour paper that they should run with the smaller sRGB gamut. Why should our professional images be compromised by this. Many supply a set-up disk to have you spoil your monitor settings to match theirs, and then re-work your file to suit. All because they have not got to grips with colour management. Perhaps I am missing something here. snip The owner of the shop agreed > that my sep was significantly better and all RGB to CMYK conversions from > the studio would now be conducted using this method. When the plant manager > asked what would be required for PrePress to do the same on their machines, > I told them all they had to do was load my CMYK ICC profile into their > copies of Photoshop. He refused saying it was "too much work" and what was > really frightening was that the owner agreed! The output was staring them in > the face but the prospect of changing a single setting in Photoshop (from > what is just a default and not a very good one) was just too scary for them > to do. To this day, the digital studio is fully color managed (and doing > more and more work I should ad) and prepress goes along in ignorant bliss > while their workflow for capture decreases. All that I can say...politely that is, is how incredibly stupid of these folk. Not to be able to place a profile into PhotoShop for considerable demonstrated benefit, this clearly demonstrates I think what many are up against in this business! No doubt you would have been happy to extend your services to include this Would you say this attitude is the exception rather than the rule in the US? I wonder if there are people across the pond in the EU with a view on this? Regards Richard Richard Kenward Richard Kenward Digital Imaging +44 (0)1873 890670 ------------------------------------------------------------------------------- From: Gordon Pritchard, INTERNET:Gordon_Pritchard@CreoScitex.com Date: Sun, Dec 31, 2000, 2:42 PM RE: [ColorTheory] confusing color management & color correction, was: Color settin gs and upgrades Chris Murphy wrote: "Matchprints don't work in the world of CTP where dot gain is at least 10% lower. Even Lo-Gain isn't low enough. People who move to CTP are dropping analog proofs like crazy and moving to digital proofs that can simulate a variety of output, including film based plate making as well as CTP, glossy coated paper to uncoated paper to newsprint. They are a lot more versatile and ultimately lead to fewer surprises." I would like to comment on the points that Chris makes: Dot gain in with CTP, as it is with film-based workflows, is effected through calibration. In a film-based workflow dot gain is simulated by the proofing media based on the belief that the same film used to image the proof will then be used to image the plate that will be used to print the job. Dot gain is (among other things) a function of ink density and halftone screen frequency (lpi). Printers will normally proof using a specific dot gain target media (e.g. SWOP, or commercial low-gain) with their respective ink densities and lpi. With CTP you can calibrate your plates to deliver whatever final dot gain you desire. More accurately, with CTP you can calibrate your plates to deliver whatever tonality you desire. Whether you like the tonality represented by 21% dot gain or 28% dot gain all you need tgo do is image calibrated plates that will deliver that tonality on press. Many of our customers use digital laser Matchprints for their CTP presswork. Some prefer the low gain, others use SWOP. Chris, you seem to be using the term "analog proofs" and "digital proofs" to mean halftone dot proofs vs "contone" proofs. This is not accurate, In a CTP workflow you can choose to have digital halftone proofs that corresspond to film-based proofs (i.e. image with the same dots that will appear on plate on the same device that will image the plates) or you can use off-platesetter halftone proofs, or contone inkjet-type proofs. Most shops that I am aware of try to provide the choice to their customers depending on the level of risk and liability they are willing to tolerate. best, gordo Gordon Pritchard Commercial Print Specialist CreoScitex Vancouver Canada T: 604.451.2700 ext 2870 C: 604.351.2437 gordon_pritchard@creoscitex.com http://www.creoscitex.com >Print, the original dot com< ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sun, Dec 31, 2000, 4:16 PM RE: Re: [ColorTheory] confusing color management & color correction,was: Color settings and upgrades >Would you say this attitude is the exception rather than the rule in the US? It's the general rule, with exceptions and an increasing number of exceptions. But I doubt in the printing industry there are more than 20% of total printers who can provide accurate separation information for their customers. >I wonder if there are people across the pond in the EU with a view on this? It depends. When it comes to advertising (magazine and newsprint) they have quite good process control and the use of ICC profiles for separations. Actually there are a couple of publications that are escaping my memory at the moment that REQUIRE the use of embedded ICC profiles. If your images don't have profiles embedded, they get rejected. Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sun, Dec 31, 2000, 4:07 PM RE: Re: [ColorTheory] confusing color management & color correction, was: Color settings and upgrades >I'm one of the vast majority who can't provide accurate separation >information for our new digital press. We hired one of the best pressman in >town and profiled the press, but it was useless. OK that's a little vague. I've profiled digital presses successfully to the point I wasn't even involved with the press itself. It's not uncommon to to prematurely profile a press before consistency has been validated. > When I talk with the >pressman he says he can't match Iris proofs and gives me an endless list >variables as to why presses can't be profiled; ink tack, ink viscosity, >blanket pressure and age, paper color and stock, layout of the particular >piece, and on and on. Ahh. So just because he hasn't been able to profile a press means that it simply cannot be done? I've done it and so have thousands of other people. >He has all kinds of tricks for making the job look >good and sees printing more of an art then science, Which means the press doesn't ever behave the same at any point in time. He's changine press behavior (tricks) in order to make the current job look good. It's a totally valid way of operating a press when separations aren't ideal. But it's asking for disaster if you're going to profile the press, and then subsequently expect the same press behavior. > saying prepress people >(and art directors, designers, etc) don't understand what happens when its >time to put ink on paper. Well that's pretty funny considering most presses aren't using a lithographic printing process. >So what good are these measurements if the press is running all different >stocks of paper, and press conditions change as a result of all the >variables listed above. Its easy to say the pressman just doesn't want to >adhere to process controls and standards, but hard to argue with quality >results and satisfied customers he produces. So I could hand out Photoshop >sep profiles, but he's always going to have to "fix" it on the press anyway. There are limits to the ability of a pressman to fix something on press. It's not possible to fix anything. The separation table (profile) must produce reasonable separations for the intended form of output. In order of increasing complexity: 1.) One profile based on the contract proof. 2.) One profile for the most commonly used stock (or two); the rest are fixed on press using conventional methods. 3.) One profile based on average press performance. 4.) One profile based on each classification of paper stocks (grouping paper stocks) 5.) One profile for each paper stock. Needless to say #5 is extremely rare. #1 is rare enough as it is. Some printing processes do not lend themselves well to correcting bad seps once they are being printed. Either they get separated correctly or they print wrong and there's nothing but either reseparation or manual correction of the file that will solve the problem. And ultimately if you're happy with your printing situation, then ignore what I'm saying. But one of two things will happen as we see customers originating images for which there is no physical original. Only a digital file and an embedded profile. 1.) They will insist on accurate separation information because compromise (fixing it on press) while acceptable to some won't be acceptable to others. You can't fix all kinds of bad seps on press. 2.) Color expectations will go down. Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sun, Dec 31, 2000, 4:51 PM RE: Re: [ColorTheory] confusing color management & color correction, was: Color settings and upgrades >Which is still this case at most printers, coast to coast and overseas. I disagree that it's as much of a problem in mainland Europe. They have older equipment yet better consistency than here because they have better process control. Adoption of ICC profiles is also higher. They are much more willing to experiment with new techniques and technology (within limits of course) than American counterparts. >But the more complex the workflow, the greater the value a CMS provides. To >benefit in that situation you need to be more than a newbie with a >SpectroLino. This argument is really out of scope because you keep bringing in more issues instead of staying on one issue. You're saying it's complicated, and I'm saying it doesn't have to be complicated for everyone, just those who already have complicated situations. >No, the original context of the argument was that I made a recommendation to >someone (new to the issues) to learn the numbers... or make a full effort to >learn about CMS. Why is there always an either or with this topic? They are totally different things. There is no OR because they aren't the same thing. >The entire point of my recommendation to Gary was that a CMS is only as good >as the knowledge of person setting it up. I wasn't even knocking Managed >Color... but I was making it clear that it's more than just buying some toys >and pushing a button. Well that's the case with every technology. You don't see people going to a CreoScitex or Heidelberg brick and mortar store, buying a complete CTP system in a box and setting it up themselves from scratch. >> Again, it depends on the situation. If you're a graphic designer, then >> you really only need to know two things about color management: >> >> 1.) The procedure for calibrating and profiling the monitor. >> >> 2.) The settings in Photoshop that you should be using, most important of >> which is getting the most accurate CMYK information plugged in. > >This is what bugs me about your line of reasoning. You're coming perilously >close to saying that the user doesn't really have to understand color; >neither the scientific nor psychological aspects. Don't put words into my mouth that I haven't said. It's rude. I never said nor implied the user doesn't have to understand color. There is a clear disclaimer above that says all you really need to know about color management are two things. I didn't say everything you need to know about color reproduction. I've been quite clear in saying that color correction (which includes a knowledge of color concepts, and basic ink on paper properties) is necessary knowlege. But color correction and color management are two different things. >> This would take at most couple of hours of training for someone who is >> already suitably experienced with offset printing. It's not that big of a >> deal. > >Not a big deal if they're just going to SWOP or sheet-feed offset... but >again, as the workflow diversifies, the knowledge needed to manage the CMS >increases. I don't understand the relevance of this argument at all. It's a "but...but...but" argument. There is no but. I qualified the statement as stating the two things the individual graphic designer needs to know about color management. Everything else is optional and dependent on their specific workflow, obviously. Saying that as the workflow gets more complicated, their knowledge about CMS needs to go up is stating the obvious. >Nice to hear you say that. The whole point of my recommendation in my >original post was that you need to know both. And yet you never said this in your original post. My first post stated your implication that color correction and color management were interchangable; that color correction and good understanding of color could supplant color management. From your 3rd post: >I'm reasonably confident that a >competent numbers guy can hit any device (that behaves according to man's >present understanding of color) with two tries. The idea a person can see only two prints from some random output device and instantly size it up is beyond human ability. It also confuses color correction (fixing images) and color management (mitigating different behavior of output devices). Now I've used up my three emails for the day. Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sun, Dec 31, 2000, 4:22 PM RE: Re: [ColorTheory] confusing color management & color correction, was: Color settin gs and upgrades Gordon writes: >With CTP you can calibrate your plates to deliver whatever final dot gain >you desire. More accurately, with CTP you can calibrate your plates to >deliver whatever tonality you desire. Whether you like the tonality >represented by 21% dot gain or 28% dot gain all you need tgo do is image >calibrated plates that will deliver that tonality on press. For some this acceptable and for others it isn't. The blurred edge of a halftone dot through the making of an analog proof (and subsequently making a plate) doesn't happen with CTP. They have sharp edge dots. In terms of color you can compensate for this, but you won't get an exact dot for dot proof. In distributed advertising scenarios, yes this is the way it has to be done because some customers are going to supply Matchprints, others will supply Waterproofs, others will supply digital proofs of some kind. All must go on press and be "matched" as close as possible. So the use of transfer curves here is important. >Chris, you seem to be using the term "analog proofs" and "digital proofs" to >mean halftone dot proofs vs "contone" proofs. Not at all. I'm using the term digital proof to mean "not made with film". Nothing more or less. For some, the lack of dots isn't a problem and they are only concerned with color. For others, they will want halftone dots. This doesn't negate the fact both can be a digital proof. Chris Murphy ------------------------------------------------------------------------------- From: "richardkenward", INTERNET:richardkenward@onetel.net.uk Date: Sun, Dec 31, 2000, 5:49 PM RE: Re: [ColorTheory] [colortheory] Color settings and upgrades Dear James..... and anyone else still interested in this thread I want to just take us back to your original posting. Ugh! What really puzzles me about your argument is this proposition that one can adjust colours according to the numbers and you can then expect to obtain predictable/accurate colour on press. a) What source are you using for the numbers? b) What is the relevance of these number with different presses, different operators, different paper stocks, different screens, different inks, different image setters, different....well I think you get my drift! As I see it, the 'numbers' are no real guide to what you will get on paper unless these numbers are VERY closely tied to a particular set of conditions. And obviously this output device must be run at a high degree of accuracy or again the 'numbers' will be of little value. I contend that in this case you have an ideal situation for a profiled press! Best regards Richard Richard Kenward Richard Kenward Digital Imaging +44 (0)1873 890670 ------------------------------------------------------------------------------- From: James Steincamp, INTERNET:steiny@cavtel.net Date: Sun, Dec 31, 2000, 5:51 PM RE: Re: [ColorTheory] confusing color management & color correction, was: Color settings and upgrades > Don't put words into my mouth that I haven't said. It's rude. Excuse me... you are the one who has maintained a consistently condescending tone throughout your messages, failed to understand what I am saying and have had little problem with putting words in my mouth. This issues that you have chosen to visit over and over are issues that were not present in my initial post and are tangential to my original line of reasoning. You maintain that I have somehow managed to confuse the idea of color correction with color management, which is nowhere present in the original message. > Then there is profiling... which in a Managed Color environment typically > means to create a ColorSync (or ICC) profile. Profiling is an acknowledgement > that no device is perfect (even when calibrated) and serves as an attempt to > describe a device's (your monitor in this case) imperfections.