To all those that think I jumped down Clayton's throat too aggressively and that it was a personal attack I apologise. But I was reacting in particular to his leading statements (repeated below). As I noted, this was not at all what I was saying. Any thoughtful, measured reading of what I wrote would not lead one to that conclusion. Nonetheless I should not have been quite so aggressive. (If you can't be bothered reading this long post but print from DG20 or any workspace that isn't GG2.2 and are trying Epson Adv B&W then at least jump to the last two paragraphs.) For those who seem to use any such post as a chance to jump on the aesthetic vs technical bandwagon you are gravely mistaken. Technical knowledge does not in any way displace aesthetics in making art (although not everyone printing B&W need be an "artist"). Technical skill does, however, help one to express their creative mind. With greater technical skill you are much more likely to achieve your artistic expression to your satisfaction. Measurements also allow us to communicate with each other without ambiguity. They say nothing about "art" or "artistic ability" and do not hinder them either. This isn't about who is smarter. (I'm sure Michael is far smarter than me - he can _read_ Aristotle.) The concepts here aren't even very difficult. It's about taking the time to understand them and to then benefit from that understanding. It is helpful to know what it means to work in one workspace vs another. The colour associated with, say, 90% K changes according to the workspace that gives it meaning. Look at a step wedge on your display - at the 90% patch. Assign GG2.2 to it. Then assign GG1.8 or DG20. The colours change because those same numbers are being given a new meaning, a new colour space for interpretation. In all cases 90% K stays 90% K. (If you "convert" rather than assign then the number changes so that the colour remains the same.) Printers only react to numbers - they have no knowledge of human vision and perception. We work visually (at least most of us do most of the time - an eye on the info palette is very helpful and people working in areas where exact colour matching is more important, eg advertising, are going to be more inclined to work with the numbers on more occasions). But we send numbers to the printer. Understanding, even at the most basic level, that we need to send the right number to the printer to get the right colour is helpful. The reaction of Epson's Adv B&W to, say, 90% K for any given workspace is not up to you. Epson programmed the printer to print a shade of grey for each possible 8 bit input value. It has a response to each possible input number and that is it. It's useful to understand what that response is. Epson did provide some controls for tweaking the response of the printer to each number but those controls are relatively crude compared with the controls we have in an image editor like PS. In general you are likely better off editing the image in the editor and gaining more precise control. Which parts of your image you decide to place at 90% K or any other % K in your chosen workspace is entirely up to your artistic desire (or where you're told to put it if you are working in an environment which doesn't provide that freedom - a corporate logo may prescribe the shade of grey used, for example). Colour management helps us a lot by providing a translation mechanism so that we need not worry about the numbers as much. Whatever parts of our GG2.2 image that we have subjectively placed at 90% K (as defined by GG2.2) can be converted to the right colour when we change colour space. Colour management also has protocols for dealing with the many situations when a colour or shade of grey can't be rendered in space to which we would like to convert so that results are still pleasing. Colour management is a general case transformation tool - it's intended to be able to handle any and all colour spaces. Epson's Adv B&W doesn't use colour management but there is nothing to say that Epson didn't think about (the narrow case and) the reaction of their printer to each possible input for ONE given workspace. In fact it is the far more intelligent thing to do than making it linear to L* (for example). They could have, as I suggested, decided that the most common colour space in use out there today is Adobe RGB and then by default the most common grey space is likely GG2.2. The logical thing to do then is to tailor the response of the printer (which they can do with a remarkable degree of accuracy for their inks and papers) to produce visually pleasing results (say, by using the same greyscale axis transforms they use in their perceptual colour profiles) for all those many people working in the most popular workspace today. If they did so, one would expect 95% K to be printed very close to 100% K (or if you don't like the absolute black then think 94 and 99) because it's meant to be that way for someone working in Adobe RGB/GG2.2. The point of this conversation was to say don't judge whether the printer is bunching shadows by printing a step wedge and looking at the results. If Epson Adv B&W's default settings have been "calibrated" for Adobe RGB/GG 2.2 then you'd expect the 90, 95 and 100 patches to be a lot closer together than if it had been calibrated for DG20. As I said above, Epson Adv B&W does not use colour management and its response is quite likely calibrated for one workspace. (It may be or may not be but I'll bet that the thought of DG20 was never entertained for a nanosecond if at all.) QTR Create ICC does, though, allow us to plug this missing gap in Epson's B&W solution. It allows us to profile the Adv B&W response (or any other workflow) and to put a general case translation device between our workspace and the response of the printer (the print space). When we use an ICC profile the numbers are converted or changed so that the printer gets the right number to produce the right colours. This is very helpful because it allows separation between us and the numbers. We need focus less on what colour is generated by Adv B&W for the 8 bit equivalent of 90% K because colour management will translate the GG2.2 or DG20 number to the right number in the print space. If you can use QTR Create ICC I highly recommend it. Roy has made it very flexible with respect to data input and you need only have a densitometer. (A spectrophotometer will give you the added ability to soft proof hue.) If you can't use QTR Create ICC then it's likely good advice to ensure that you work in Adobe RGB/GG 2.2 when using Epson Adv B&W. Of course you can still decide that you don't like the default "Darker" setting - that's entirely up to you. But make that judgement not by looking at a printed step wedge (in which case one would likely expect the black patches to be bunched at the end) but rather from the results of prints printed from the GG2.2 workspace - the workspace for which Adv B&W was most likely calibrated. (Remember if you are not using colour management your choice of workspace is very very important.) If you use a workspace like DG20 (without colour management) then it is not at all surprising that you wouldn't like the response of Epson's Adv B&W default settings. The reason is obvious with a little technical knowledge. Look at a step wedge on your display. Assign DG20 to it. Then assign GG2.2. Those same numbers produce much darker 80-95 patches, for example, in GG2.2 than in DG20. The numbers are what go to the printer not the colours you see. (Without colour management there is no translator to convert the numbers so that the colours don't change.) Of course the prints will come out a lot darker than on your display and your shadows will look blocked up. But the reason is your lack of use of colour management (now available via QTR Create ICC) and, given that, your poor choice of a workspace which doesn't align well with how Epson likely calibrated the response of the printer! The table below is useful in making this point clearer. It is the lightness (as measured by CIE Lab's L*) ascribed to 80, 85, 90, 95 and 100 K for each of DG20, GG1.8 and GG 2.2: Workspace 80 85 90 95 100 DG20 30 23 15 5 0 GG1.8 28 21 13 4 0 GG2.2 20 13 6 1 0 So if you look at a patch (or part of your image) that's 85 K in DG20 but print it with something "calibrated" to Adobe RGB/GG2.2 you can see it's going to print a lot darker than it looks on screen and furthermore the gap between it and a 95K patch will be a lot closer than you expect. There's nothing wrong with the printer - it's just reacting to the same number. You just need to take the time to understand why there's a difference between what you see on screen and what you see in print and amend your practice to suit. Anyone with an image tagged with DG20 (or any other non GG 2.2 space) now looking to use Epson Adv B&W would likely be well advised to convert that image file to GG 2.2 prior to printing it (firstly with the default settings). That's a far more sensible first step than trying to tweak the driver settings. > From: Clayton Jones <cj@...> > > So according to what you are saying, photographers who now do digital > printing aren't supposed to bring their personal aesthetic judgments > to the work any more. Instead we're supposed to plug in the numbers > and accept what comes out because someone somwhere decided that GG2.2 > is what we're supposed to use and that dark values are supposed to be > bunched up. Hogwash!
Message
Re: [Digital BW] Re: Comparison: K3 versus Ultrachrome inks on Semi-Matte + ImagePri
2005-11-21 by Steve Kale
Attachments
- No local attachments were found for this message.