From: INTERNET:Iganatz@aol.com, INTERNET:Iganatz@aol.com Date: Fri, Dec 8, 2000, 3:12 PM RE: [ColorTheory] 6.0 Profile to Profile conversions Hello all, Loving 6.0 so far, with the exception of profile to profile conversions. Most of the time, I prefer to get my hands dirty and generate seperations via PS's built-in engine, but there are times when some of those LinoColor profiles come in mighty handy (til PS6 anyway). The profile to profile setup: RGB working space is custom calibrated monitor profile (same as I used in 5.5), CMYK work space is one of Heidleberg's Offset SWOP profiles (same as 5.5). When converting, source profile and destination profile are exactly the same as those I used in 5.5 (again, LinoColor profiles out of the box). Still using the same engine for said conversions, same settings, etc. Absolutely no change in hardware, software, OS. When I do a profile to profile conversion, and then perform the same conversion in PS 5.5 using the same parameters, the results differ. This wouldn't be a problem (as the new seperations look just fine), but on certain imagery, I need to be absolutely certain that the seperations are identical. So what am I missing? Is it a bug, my own error, or a change in PS's color architecture? Anyone have the same problem? Experiences to relate? Workarounds? TIA, adam sanders P.S. All hail the sRGB workspace! (Kidding. sorry I said anything, mumble, mutter...) ------------------------------------------------------------------------------- From: "shAf", INTERNET:michael@shaffer.net Date: Fri, Dec 8, 2000, 4:01 PM RE: RE: [ColorTheory] 6.0 Profile to Profile conversions adam writes ... > Loving 6.0 so far, with the exception of profile to > profile conversions. > ... > When I do a profile to profile conversion, and then > perform the same conversion in PS 5.5 using the same > parameters, the results differ. .. How are the results differing ... as presented on the monitor? ... or as printed? PS6 has changed in the way it presents differing color spaces on the monitor. With PS5, all images are shown in your preferred working space, and only that space. With PS6, images are now shown in there own color space ... i.e., whatever you p2p'd to. shAf :o) ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Fri, Dec 8, 2000, 6:09 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions on 12/8/00 2:11 PM, Iganatz@aol.com at Iganatz@aol.com wrote: > The profile to profile setup: RGB working space is custom calibrated > monitor profile (same as I used in 5.5) Bad dog! You never use a display profile for a Working Space. Display profiles are only used for one function, to preview images. Use a Working Space (Adobe RGB perhaps). > When I do a profile to profile conversion, and then perform the same > conversion in PS 5.5 using the same parameters, the results differ. It shouldn't (if it did, it would be very minor due to a slight difference in how Photoshop 6 deals with a rendering intent to the screen, again depending on your settings). Here's probably the issue. You don't have that Heidelberg profile loaded in your CMYK Set-Up in Photoshop 5. In Photoshop 6 when you do a convert to profile, that profile in that document is used for the preview because Photoshop 6 supports documents specific color. Photoshop 5 doesn't. So if you have another profile loaded in CMYK Set-Up in Photoshop 5 and do a Profile to Profile, THAT other profile is being used for the preview. Andrew Rodney ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sat, Dec 9, 2000, 12:38 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions >I use ColorMatch as my RGB editing color space because it is pretty close to >Fuji Pictro output but I am also tempted to use my monitor RGB color space >as >it is even closer. That's really surprising as I haven't seen a monitor and a Fuji Pictro that had the same behavior such that an RGB image viewed with something like Picture Viewer (which does NOT compensate for the monitor) would print out the same on a Pictro. They are very different devices so I wonder if this is what's really happening or if it just appears that way. In any event, if the only output you ever do is to the Fuji Pictro then you have a better case for using the Pictro's profile as the working space, not your monitor profile; and even here I would disagree with this workflow because Pictrography's don't exhibit good gray balance and therefore neither would any working space based on a Pictro. Color correcting images in color spaces based on devices (like monitors, scanners, cameras or RGB output devices) is just not an ideal situation. You lose several of the major benefits of working in RGB in the first place. > If I want to share my RGB file I can convert to any other >available color spaces. The less often you convert the better. You are better off standardizing on one space such as ColorMatch RGB and using that for color correction and archiving. When outputting, all derivative images come from that master file. Standardizing on a non-standard space, let alone one that is also a device space just isn't a good idea. Will it be the end of the world? Will your images suck? Definitely not. But there is a greater benefit of using a color space based off a well defined mathematical model instead of the very inconsistent and non-linear behavior we get from devices (and therefore color spaces based on devices). > If I want to keep using ColorMatch (or any other >recomended) I would have to convert each file to custom created fuji profile >prior to printing. Yes that's the workflow. You color correct and archive in a well behaved space, and then convert to device spaces when you are going to output to a device. > What is exactly bad in working in good monitor >space? They aren't uniform and they aren't gray balanced. i.e. there is no such thing as a good monitor space. None of them are uniform and none of them are gray balanced. That's the nature of devices. Only mathematical color spaces have uniformity and gray balance. > I think ColorMatch is a color space based on Radius PressView >monitor >and before Photoshop 5 people were working in their monitor spaces and >although did not have such flexibility in sharing RGB files like with >current >ICC workflow, color of output was still good. In regards to ColorMatch and PressViews, Radius had a serious amount of proprietary technology built into those monitors to force them into a very specific non-device like behavior which required constant attention in order to maintain. The internal circuitry was making constant adjustments (not nearly to the degree of a Barco, but whole bunch more than a regular desktop monitor). Daily calibrations were necessary to maintain this exactly. So it's a special case device, so is the Barco. As for before Photoshop 5 this gets really complicated because Photoshop actually didn't use monitor RGB in the sense that Photoshop 6 allows. Photoshop 6 allows actual monitor behavior to be used to define a color space, including non-uniformity and lack of gray balance. Photoshop 2.5, 3, and 4 for example, didn't do this. You could specify your monitor model's PRIMARIES, a single gamma value, and a white point and that was it. This assured that the resulting RGB space used for editing was well behaved, uniform and gray balanced. >One question that pops >in >my mind is why most digital photography labs labs use relatively small gamut >RGB spaces - AppleRGB (Burrel Lab), sRGB (Miller Lab) and maybe the answer >is >that because those color spaces are very close to standard monitor spaces >and >printers that print with light instead of ink like those spaces better? You'd have to ask them. I don't see the benefit of using a smaller space with larger gamut output devices that are laser diode and CRT based. I would use something like Adobe RGB for all of them. It it a good space for output to press as well as one of these large gamut output devices (and large format printing as well for that matter). Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Fri, Dec 8, 2000, 10:46 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions >> The profile to profile setup: RGB working space is custom calibrated >> monitor profile (same as I used in 5.5) > >Bad dog! You never use a display profile for a Working Space. Display >profiles are only used for one function, to preview images. Use a Working >Space (Adobe RGB perhaps). Well you can with Photoshop 6 which I really don't like. I was going to say something but didn't because I don't know enough about his workflow. If the workflow is really closed loop, then there is probably nothing severely wrong with using a monitor space (or digital camera space) as an editing space, but it's definitely NOT ideal. I would definitely recommend using ColorMatch RGB or Adobe RGB (if you can't decide, flip a coin) for work destined for process printing and not use the monitor profile as a working space. The monitor profile is used only for Display Compensation which Photoshop does behind the scenes based on the monitor profile set in the ColorSync or Monitors control panel (depending on the version of ColorSync you have). There is no need to use the monitor profile as a working space. And if you share RGB images it's a really bad idea to use a monitor profile as an RGB working space. >Here's probably the issue. You don't have that Heidelberg profile loaded in >your CMYK Set-Up in Photoshop 5. I think what we need to find out is if the differences are visual (on-screen) or if the differences are numeric with the actual separation. I was assuming the numbers were different. Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Fri, Dec 8, 2000, 10:40 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions > When I do a profile to profile conversion, and then perform the same >conversion in PS 5.5 using the same parameters, the results differ. By how much? There are two possible reasons for this. First, the Adobe engine is different and you will get different results using it regardless of whether you use ICC profiles or not because there is no such thing as not using an ICC profile in Photoshop 6. All profiles are ICC. If you load a separation table, Photoshop builds an ICC profile and then uses that. Second, the black point compensation algorithm has been improved. Third, by default all conversions on 8-bit/channel images have a dither (noise) added to them in order to prevent posterization. These would explain differences of MAYBE a few percent. If your differences are greater then it's something else. > This >wouldn't be a problem (as the new seperations look just fine), but on >certain >imagery, I need to be absolutely certain that the seperations are identical. Not going to happen. It's impossible to use the old Adobe engine in Photoshop 6 or use the new Adobe engine in Photoshop 5. So there will likely always be an occasional pixel with a 1%-3% difference between the versions; probably closer to 1% if you were to turn off the dither feature. Chris Murphy ------------------------------------------------------------------------------- From: INTERNET:rproulx@r2p2.com, INTERNET:rproulx@r2p2.com Date: Sat, Dec 9, 2000, 9:43 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions On 8 Dec 2000, at 23:37, Chris Murphy wrote: > >I use ColorMatch as my RGB editing color space because it is pretty > >close to Fuji Pictro output but I am also tempted to use my monitor > >RGB color space as it is even closer. > > That's really surprising as I haven't seen a monitor and a Fuji Pictro > that had the same behavior such that an RGB image viewed with > something like Picture Viewer (which does NOT compensate for the > monitor) would print out the same on a Pictro. They are very different > devices so I wonder if this is what's really happening or if it just > appears that way. I've been working to profile a few "direct to photo paper" printers and have struggled to grasp this whole RGB colorspace thing. I do sense some light at the end of the tunnel. My current conclusion is that a photo print's color gamut can be contained within any RGB space offered by Photoshop. If I have an image in a very wide space such as Kodak's ProPhotoRGB and do a profile to profile conversion to the Pictro I see no difference in the output between this and converting the image to sRGB and doing the conversion from there. Like offset printing, photo prints have a very limited color gamut (much smaller than transparency film) and the colors it can reproduce will fit in even a crippled RGB space such as sRGB. Kodak even arms photographers to defend themselves against pesky clients (such as fine art painters) who demand that we match their colors. http://www.kodak.com/global/en/professional/support/techPubs/e73/e73.shtml If we only have 3 crayons to play with...we can only ceate so many colors. It's 3 limited photo emulsion layers and not agile color guns that we're playing with. One huge mistake I was initally making was to view the image in other colorspaces without converting it to that space first. Big mistake...I didn't understand the connection between a colorspace and it's profile. This is my "current " conclusion. Though I'm always hoping to learn where it's wrong PS: Bought Dan's new PS6 book. Couldn't resist. It's -20C (-4F) outside so it should keep me occupied until these tulips start popping up. Russell Proulx Photographer Montreal, CANADA ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Sat, Dec 9, 2000, 10:34 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions on 12/9/00 8:43 AM, Russell Proulx at rproulx@r2p2.com wrote: > I've been working to profile a few "direct to photo paper" printers and have > struggled to grasp this whole RGB colorspace thing. I do sense some light > at the end of the tunnel. My current conclusion is that a photo print's color > gamut can be contained within any RGB space offered by Photoshop. Not so. I can send you a gamut map of a Pictrography 3000 or 4000 or a Lambda laid on top of sRGB and there are areas that fall outside sRGB (and areas of sRGB that are wider than these Working Spaces) > If I have an image in a very wide space such as Kodak's ProPhotoRGB and > do a profile to profile conversion to the Pictro I see no difference in the > output between this and converting the image to sRGB and doing the > conversion from there. I'd agree that the differences (depending on the image itself) is either very subtle or nonexistent. The key here is the original image. I do have some scans where the gamut is so wide that there are some areas where you do see increased saturation in the wider working space. For all but the most anal end user, the differences are not that striking. > Like offset printing, photo prints have a very limited > color gamut (much smaller than transparency film) and the colors it can > reproduce will fit in even a crippled RGB space such as sRGB. SRGB at best will give you about 85% cyan. If you don't care about the other 15%... The other issue with starting with a smaller Working Space is you hose whatever increased color gamut you started with when converting into that space. It's quite possible that a year from now, you'll want that wider gamut data when someone has a wider gamut output device to take advantage of it. Why toss the data for good in an RGB archive? > One huge mistake I was initally making was to view the image in other > colorspaces without converting it to that space first. Big mistake...I didn't > understand the connection between a colorspace and it's profile. You just illustrated a good lesson in how the new "Assign Profile" command in Photoshop 6 works. Assigning a profile to a file not in that colorspace alters the recipe for Photoshop and so the preview changes. It illustrates how a file is nothing more then a series of numbers. The numbers mean nothing without a description (the profile). There was an interesting argument on the Adobe Photoshop forum from someone who really believed that 70C 70M 80Y 70K really meant something. He believed those numbers expressed a color somehow. You just saw how assigning a profile in Photoshop 6 with the same numbers change the meaning of the colors. Andrew Rodney ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sat, Dec 9, 2000, 10:59 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions > My current conclusion is that a photo print's >color >gamut can be contained within any RGB space offered by Photoshop. That would be an inaccurate conclusion. You might conclude that a lot of the images you print don't have lots of vivid saturated colors and therefore sRGB v. ProPhoto RGB yields little difference. The gamut of a Pictrography, Lambda, or Lightjet on reflective media is larger than sRGB, it's just that you're probably not noticing it with most images you're using, and in fact it might be that most images in general don't stretch the limits of these devices anyway. >If we only have 3 crayons to play with...we can only ceate so many >colors. It's 3 limited photo emulsion layers and not agile color guns that >we're playing with. Well it's not so much about the number of crayons but the purity of the color they can reproduce. If it were possible to create dyes (or pigments) of 100% purity, then you would only need three. Theoretically. Chris Murphy ------------------------------------------------------------------------------- From: INTERNET:rproulx@r2p2.com, INTERNET:rproulx@r2p2.com Date: Sat, Dec 9, 2000, 11:31 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions On 9 Dec 2000, at 9:58, Chris Murphy wrote: > >A photo print's color gamut can be contained within > > any RGB space offered by Photoshop. > > That would be an inaccurate conclusion ....it might be that most images in general > don't stretch the limits of these devices anyway. Thanks for setting me straight (Andrew did as well) :-) > Well it's not so much about the number of crayons but the purity of > the color they can reproduce. If it were possible to create dyes (or > pigments) of 100% purity, then you would only need three. > Theoretically. I never meant to suggest that it's not theoretically possible to recreate the whole visible color spectrum with 3 colours. I was speaking more in real world terms and the fact that there's currently no such magic photo emulsion. The color gamut of currently available color photo paper is far from perfect and colours will also vary according to brand, type, chemistry quality and type, and a pile of other things. You were right mention that color theory suggests that such perfection is theoretically possible. Russell Proulx Photographer Montreal, CANADA ------------------------------------------------------------------------------- From: INTERNET:rproulx@r2p2.com, INTERNET:rproulx@r2p2.com Date: Sat, Dec 9, 2000, 11:09 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions On 9 Dec 2000, at 9:32, Andrew Rodney wrote: > I'd agree that the differences (depending on the image itself) is > either very subtle or nonexistent.....For all but the most anal > end user, the differences are not that striking. I stand corrected. But I couldn't see much of a difference "with my eyes". I began doing this comparison because of an email I received from someone who was told by a lab that their Pictrography color space matched that of sRGB. That didn't sound right. Yet when I converted from one RGB space to another with an image of a Kodak refective target I couldn't see any difference on screen. Measuring various neutral patches in the "converted to printer profile" version seemed also to show no difference (ie: the same numbers would be sent to the printer). > The other issue with starting with a smaller Working Space is you hose > whatever increased color gamut you started with when converting into > that space. I personnaly would not want to use sRGB as a working space. I'm even tempted to play with Kodak's ProPhoto RGB for a while and see what it has to offer (when working with images destined for photo output only). > There was an interesting argument on the Adobe Photoshop forum from > someone who really believed that 70C 70M 80Y 70K really meant > something. He believed those numbers expressed a color somehow. I see that as a SWOP color with a very slight orange tinge to what's already very black. Interpreting those numbers is easier than interpreting an RGB value without knowing it's RGB workingspace. Then again, without knowing the details of the press and paper it's really not all that much different....ie: it's only one piece of a bigger puzzle that needs all the parts identified before we can form any conclusions. Gee Andrew...we're beginning to agree Russell Proulx Photographer Montreal, CANADA ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Sat, Dec 9, 2000, 11:55 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions on 12/9/00 10:09 AM, Russell Proulx at rproulx@r2p2.com wrote: > Yet when I converted from > one RGB space to another with an image of a Kodak refective target I > couldn't see any difference on screen. And you wouldn't. You'd have to output the file or in Photoshop 6 use the Convert to Profile command with each and either visually examine both files in OUTPUT space or subtract one from the other. > I personnaly would not want to use sRGB as a working space. I'm even > tempted to play with Kodak's ProPhoto RGB for a while and see what it > has to offer (when working with images destined for photo output only). Two extremes of the spectrum. Adobe RGB 1998 would be safe space for most all output needs. But again, who knows what we will be printing with in 5 years. > Interpreting those numbers is easier than interpreting > an RGB value without knowing it's RGB workingspace. What's the difference? You mean because the CMYK is assigned as SWOP? I guess that's a bit more of a description of the CMYK then no profile for the RGB file but not a huge once you consider how vague "SWOP" is. Now if you're using someone's SWOP profile, then that profile is describing *some* SWOP condition. > Gee Andrew...we're beginning to agree Now I'm getting scared Andrew ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Sat, Dec 9, 2000, 11:55 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions on 12/9/00 9:58 AM, Chris Murphy at lists@colorremedies.com wrote: > The gamut of a > Pictrography, Lambda, or Lightjet on reflective media is larger than > sRGB, it's just that you're probably not noticing it with most images > you're using, and in fact it might be that most images in general don't > stretch the limits of these devices anyway. I think that's the case a lot of the time. In theory (the cause of this list) and mathematically, this is absolutely true. In practice, most of the times I've tested sending out images using the various Working Spaces to these devices, I haven't seen a really big difference between ColorMatch RGB on one end and ProPhoto on the other. Kodak did supply me with a file to test that had a huge gamut (and was so saturated it was in my opinion butt ugly) but I did see a benefit of their Working Space on output. But this image was so non-typical and I think created to show the benefits of ProPhoto off that it was interesting as an exercise in what one could do in really extreme cases. It wasn't typical by a long shot. Andrew Rodney ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sat, Dec 9, 2000, 3:13 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions >What's the difference? You mean because the CMYK is assigned as SWOP? I >guess that's a bit more of a description of the CMYK then no profile for the >RGB file but not a huge once you consider how vague "SWOP" is. Now if you're >using someone's SWOP profile, then that profile is describing *some* SWOP >condition. Just to clarify (I know Andrew knows this) SWOP. SWOP in practice is vague only because so many things claim to be SWOP but aren't. Photoshop 3 through 6 have something claimed to be "SWOP" and yet aren't really SWOP. Many printers claim to follow SWOP but don't (they follow parts, but the actual press behavior doesn't conform.) Technically SWOP is as specific as any specification on the planet at this point in time. For 938 CMYK values we have CIE L*a*b* values the define SWOP behavior. That is, if you follow SWOP, and your press or proofer or separation exhibits SWOP behavior - there are 938 CMYK values that should target predefined L*a*b* values, or it isn't really SWOP. This means for specific CMYK values printed per SWOP, there is an expected colorimetric value that can be measured; and if there is too much deviation then by definition that process is NOT SWOP. Pretty darn specific. But so few people really follow SWOP. To get a copy of the report and measurement data for these 938 CMYK values, you can go to www.npes.org, go to the standards area and you will find a section for ordering publications, one of which is TR001. That is the report that defines SWOP to specific colorimeteric performance. Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sat, Dec 9, 2000, 3:07 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions >I never meant to suggest that it's not theoretically possible to recreate >the >whole visible color spectrum with 3 colours. I was speaking more in real >world terms and the fact that there's currently no such magic photo >emulsion. The color gamut of currently available color photo paper is far >from perfect and colours will also vary according to brand, type, >chemistry quality and type, and a pile of other things. You were right >mention that color theory suggests that such perfection is theoretically >possible. Film used for making movies uses layers of colorant made of CMY. They are quite pure. So pure they are able to get black without the use of K just by combinding CMY. There are a number of thermal transfer and dye sublimation processes than use only CMY. They aren't perfectly pure so you don't get a perfect black with equal amounts of CMY; but through the use of some amount of CMY it's possible to get black without a separate black colorant. So it's possible to get CLOSE to purity, maybe the accurate term would be "pure enough". Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sat, Dec 9, 2000, 3:02 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions >> There was an interesting argument on the Adobe Photoshop forum from >> someone who really believed that 70C 70M 80Y 70K really meant >> something. He believed those numbers expressed a color somehow. > >I see that as a SWOP color with a very slight orange tinge to what's >already very black. Only if SWOP inks are being used. If those values are sent to a large format inkjet printer, it would print black. If it were a large format electrostatic printer, it would print a black with a slight green cast. The color you get depends on the device. Hence, device dependent color. > Interpreting those numbers is easier than interpreting >an RGB value without knowing it's RGB workingspace. Actually there is greater variation among CMYK devices than among RGB devices. Knowing the basis for device behavior is MUCH more important with CMYK output than RGB output. Chris Murphy ------------------------------------------------------------------------------- From: "Dave Badger", INTERNET:dbadge@worldnet.att.net Date: Sat, Dec 9, 2000, 4:12 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions > The monitor profile is used > only for Display Compensation which Photoshop does behind the scenes > based on the monitor profile set in the ColorSync or Monitors control > panel (depending on the version of ColorSync you have). Could you expand on this? I always thought that ColorSync's System Profile was the same thing as the monitor profile causing all color to be run though that profile before being displayed on screen. So I never understood what effect picking a profile in the Monitors Control panel had on Photoshop. (I'm using ColorSync 2.6.1 with Sys 8.6) Dave Badger ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sat, Dec 9, 2000, 4:33 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions >Could you expand on this? I always thought that ColorSync's System Profile >was the same thing as the monitor profile causing all color to be run though >that profile before being displayed on screen. So I never understood what >effect picking a profile in the Monitors Control panel had on Photoshop. >(I'm using ColorSync 2.6.1 with Sys 8.6) Probably not without confusing the issue. There are actually two distinct issues: 1.) ColorSync 2.6.1 and earlier, and ColorSync 3.0 and later. All versions of ColorSync up to version 2.6.1 had a control panel where you would set the System Profile. The System Profile is what Photoshop would use as the basis for Display Using Monitor Compensation, therefore we would set the System Profile to our display profile. With ColorSync 3.0 and higher, the control panel no longer has a System Profile option, but rather a Display Profile. The catch is that you can't change it in the ColorSync control panel, you can only select it in the Monitor control panel. A new feature of ColorSync 3.0 that this provides, is that each display connected to a Macintosh has it's own Monitor control panel WINDOW so you can select a profile for each display. Previously, if you had more than one monitor connected, only one profile could be specified. The second one is so confusing I recommend MOST people not even read this. Sometimes it's better NOT to know things: 2.) Confusing the distinction between calibration and characterization (profiling). We can thank Apple for this. Calibration inherently means changing device behavior. We do this to make it consistent. From one day to the next, device behavior drifts and calibration allows us to change the device behavior back to a reference point. That's all it does. It's not for making a monitor act like a press. It's only for consistency. It's tells us nothing about what kinds of colors a device can produce. Characterization is about recording device behavior into a file. It tells us about what kinds of colors a device can produce. This means both what color a value represents, and what value is needed to get a specific color (e.g. 40% cyan looks like what on device a? What CMYK/RGB values are needed to produce a specific reddish-orange on device a?) Profiles don't change device behavior. They just provide a means to describe how that device behaves and records it in a text file. ICC profiles historically contain only characterization information, hence profiles. They are the result of profiling. Apple, in it's infinite wisdom allowed for tags in ICC profiles to record monitor calibration data as well. So this makes monitor ICC profile both a means of characerization and calibration. There is a reason why Apple did this: It was possible to have a monitor calibrated to some behavior, like a 5000K whitepoint and a 1.8 gamma, but to select a profile for that display that said the monitor had a 6500K whitepoint and 2.2 gamma. Now what happens with ColorSync 3.0 is when you go to the Monitors control panel and select a monitor profile, the operating system reads the calibration information in the profile and sends that to the video card thereby (hopefully) making the monitor behave per the white point and gamma the profile describes while at the same time selecting that profile. So you get more of an assurance that both calibration and the profile are set at the same time. This applies ONLY to monitor profiles by the way. It also explains why your monitor behavior visibly changes if you click on different profiles in the Monitors control panel (sometimes you'll click on one and nothing happens, this is an older profile that doesn't contain calibration information). Chris Murphy ------------------------------------------------------------------------------- From: Bob Smith, INTERNET:rmsmith@calpha.com Date: Sat, Dec 9, 2000, 4:52 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions Dave Badger wrote: > I always thought that ColorSync's System Profile > was the same thing as the monitor profile causing all color to be run though > that profile before being displayed on screen. correct. > So I never understood what > effect picking a profile in the Monitors Control panel had on Photoshop. > (I'm using ColorSync 2.6.1 with Sys 8.6) In your setup, picking a profile within the Monitors control panel should be avoided unless you're using the built in visual calibrator there. Picking a profile there will load the LUT for that profile and may over ride a calibration done elsewhere. In your setup you set the System profile by choosing it (your monitor profile) in the ColorSync control panel. In ColorSync 3, the System Profile is gone and you see a Display Profile choice (among profiles for other devices as well) however you cannot set that profile from the that control panel. The profile that appears there is the one selected in the Monitors control panel. Photoshop gets its info about display condition from the ColorSync control panel so it depends on the version of ColorSync that you're using where you actually make that selection. Bob Smith ------------------------------------------------------------------------------- From: INTERNET:rproulx@r2p2.com, INTERNET:rproulx@r2p2.com Date: Sat, Dec 9, 2000, 4:55 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions On 9 Dec 2000, at 14:06, Chris Murphy wrote: > Film used for making movies uses layers of colorant made of CMY. They > are quite pure. So pure they are able to get black without the use of > K just by combinding CMY. ...So it's possible to get CLOSE to > purity, maybe the accurate term would be "pure enough". Chris, This thread risks drifting off subject (it's becoming a photography thread) so I'm reluctant to bite....but I will anyway Virtually all photographic films use a subtractive method to reproduce color by converting RGB sensitive emulsion layers into CMY layers. It's not just reserved for the movie business, it what makes modern color film work. Getting enough DMax built up to prevent light from passing through a photographic emulsion can be done with any combination of color or substances. Black, with respects to film preventing light from passing, is hardly a measure of attaining any theoretical color purity. The true challenge is to reproduce all colors in the visible spectrum using only 3 dyes or pigments (or 4 or 5 for that matter). This is where theoretical purity runs into what's now humanly possible. If color purity with 3 colors was possible then art stores would have 3 very expensive tubes of pigment for sale. They don't because it doesn't exist at any price. Russell Proulx Photographer Montreal, CANADA ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Sat, Dec 9, 2000, 5:07 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversion >Virtually all photographic films use a subtractive method to reproduce >color by converting RGB sensitive emulsion layers into CMY layers. It's >not just reserved for the movie business, it what makes modern color film >work. OK what I should have said is they are working with substantially more pure CMY colorants than what we see in other areas, like inkjets or lithography. No implication that only movie film uses CMY colorants was intended. > >Getting enough DMax built up to prevent light from passing through a >photographic emulsion can be done with any combination of color or >substances. Yes OK that's true. I think there is a particular brand of ice cream that is licorice flavored and died black. The actual color used to die the ice cream black is green. They just use enough dark green dye that ultimately it appears black. The only reason why this works though, is that the green dye isn't made up of just cyan and yellow colorant. If it were, adding more green dye would eventually do nothing. You'd have green ice cream and adding more green dye wouldn't change the color, but would just give you less ice cream. The green eye also has some magenta colorant in it, and therefore once a certain amount of the dye is in the ice cream, the ice cream appears black (most all green light is effectively absorbed also, in addition to most red and blue) - or rather a very dark gray. >This is where theoretical >purity runs into what's now humanly possible. If color purity with 3 colors >was possible then art stores would have 3 very expensive tubes of pigment >for sale. They don't because it doesn't exist at any price. I refer to these man-made theoretical versions of pure colorants as NASA cyan, NASA, magenta, and NASA yellow. Each sells for an estimated price of $100,000 per ounce. So even if it did exist, it wouldn't matter. From a nature standpoint, if you go along with the idea that the human visual system is sensitive to red, green, and blue light, and that we can see with perfect saturation; clearly CMY only colorants that are pure exist in nature. Chris Murphy ------------------------------------------------------------------------------- From: INTERNET:rproulx@r2p2.com, INTERNET:rproulx@r2p2.com Date: Sat, Dec 9, 2000, 5:14 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions On 9 Dec 2000, at 16:07, Chris Murphy wrote: > From a nature standpoint, if you go along with the idea that the human > visual system is sensitive to red, green, and blue light, and that we > can see with perfect saturation; clearly CMY only colorants that are > pure exist in nature. It's called "God" ;-) Russell Russell Russell Proulx Photographer Montreal, CANADA ------------------------------------------------------------------------------- From: Darren Bernaerdt, INTERNET:darren@dbphoto.net Date: Sun, Dec 10, 2000, 10:46 PM RE: RE: [ColorTheory] 6.0 Profile to Profile conversions > > >One question that pops > >in > >my mind is why most digital photography labs labs use relatively small gamut > > > >RGB spaces - AppleRGB (Burrel Lab), sRGB (Miller Lab) and maybe the answer > >is > >that because those color spaces are very close to standard monitor spaces > >and > >printers that print with light instead of ink like those spaces better? > > > You'd have to ask them. I don't see the benefit of using a smaller space > with larger gamut output devices that are laser diode and CRT based. I > would use something like Adobe RGB for all of them. It it a good space > for output to press as well as one of these large gamut output devices > (and large format printing as well for that matter). > > > Chris Murphy One lab that I use has a LightJet and the front end is an WinNT 4.0 box running software that Cymbolic Sciences supplied with the printer. It is totally ICC unaware as well as NT not using ICC profiles. After alot of testing, the staff at this lab found that sRGB was the most successful space to feed it images from. Wouldn't it make sense since the software was developed for a NT box? (I'm thinking - assume an average Windows monitor, probably what the engineers were using...) Maybe this is why your lab is using a small gamut space? All I know is that an average image is looking great off the LightJet that my lab uses, so I supply them an sRGB image and keep sending the jobs out the door. Darren Bernaerdt ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Mon, Dec 11, 2000, 12:05 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions on 12/10/00 8:01 PM, Darren Bernaerdt at darren@dbphoto.net wrote: > After alot of testing, the staff > at this lab found that sRGB was the most successful space to feed it images > from. I doubt it. Now, if you're saying that of all the RGB Working Spaces in Photoshop, the best prints came out using an sRGB file... But sRGB isn't an output space, it's an editing space. I'll bet that having a custom ICC profile for that output device and then converting to that space (from sRGB or better, a slightly wider gamut space) would produce better quality output. Lot's of people and some manufacturers suggest that just sending out something in sRGB is all users need to worry about in an attempt to simplify all this color management stuff. They don't have to deal with producing a custom output profile or expecting users to convert files to that space. It would be easy to test this theory. Profile the device and compare the gamut map to that of sRGB and see if this is indeed true. I'll bet in many cases, it's not. Now, that being said, some input manufacturers (people that make lower end scanners and cameras) ARE forcing the resulting color into sRGB. They are not doing this to make the files or eventual output the best it could be but rather to funnel the color to the lowest common dominator. After all, we're talking about Microsoft in large part here. This funneling into sRGB is by and large a good idea for consumer oriented products and a really bad idea for anyone else that wants the best possible quality a device can produce. Andrew Rodney ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Mon, Dec 11, 2000, 2:36 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions > They are >not doing this to make the files or eventual output the best it could be but >rather to funnel the color to the lowest common dominator. After all, we're >talking about Microsoft in large part here. This funneling into sRGB is by >and large a good idea for consumer oriented products and a really bad idea >for anyone else that wants the best possible quality a device can produce. Yeah I think it's important for people to understand the predominant reason MicroSoft is doing this and pushing for it (heavily) for consumer devices. The reason is support costs. People don't get monitor to inkjet or scanner to monitor or scanner to inkjet matches that are reasonable and so they call the printer manufacturer and they blame it on MicroSoft or the scanner vendor. The scanner vendor blames MicroSoft and the printer vendor. So manufacturers have support costs and it's too complicated to have free tech support to explain color management. Lowest common denominator as long as everything looks reasonably similar is really ideal in that world. The problem is when some manufacturers who didn't understand what HP and MicroSoft were trying to do decided to do this with professional-consumer level products. Now those products are limited to sRGB at best. Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Mon, Dec 11, 2000, 2:31 AM RE: RE: [ColorTheory] 6.0 Profile to Profile conversions >One lab that I use has a LightJet and the front end is an WinNT 4.0 box >running >software that Cymbolic Sciences supplied with the printer. It is totally ICC >unaware as well as NT not using ICC profiles. After alot of testing, the >staff >at this lab found that sRGB was the most successful space to feed it images >from. Wouldn't it make sense since the software was developed for a NT box? For two reasons I doubt this has anything to do with sRGB. First, I've profiled a Lightjet with such a configuration and it has a gamut larger than sRGB. There's no indication that it's expecting or assuming sRGB and then doing a built-in proprietary conversion of some kind from sRGB to some other RGB needed for the LightJet. It takes the RGB values you give it (in the digital file) and outputs them raw except for a linearization transform function. What's probably happening is that the lab feels they get closer output compared to what they see on a monitor when Photoshop is using sRGB instead of one of the more saturated spaces because those spaces are defined by a matrix. You have just three primaries and in terms of comparing those primaries to output, they are a lot more saturated and probably not printable. So what I'm saying is that sRGB looks better in comparison to using other working spaces only because what SHOULD happen (convert to Lightjet RGB before printing) isn't happening. It's like comparing a car running on alcohol to no fuel at all when it was designed to run on conventional gasoline. Yeah it runs better than without fuel in the sense that it will run, but it doesn't run very well on alcohol because it wasn't designed that way. >(I'm thinking - assume an average Windows monitor, probably what the >engineers >were using...) I think it has more to do with the fact that despite being a larger gamut RGB device, the Lightjet color space is not defined by primaries. Working spaces are defined by primaries. This means 100% red, green, blue, or 100% red+green, green+blue or red+blue are a *lot* more saturated than what you get on any output device including the lightjet. Hence, the lower saturated primaries of sRGB makes it seem like it's a more appropriate space. In reality *NO* matrix profile can adequately define the behavior of a lightjet. You need a table based profile to do this. > All I know is >that an >average image is looking great off the LightJet that my lab uses, so I >supply >them an sRGB image and keep sending the jobs out the door. Ultimately you use what works. I assume they aren't using ICC profiles themselves. If you were to make one, you would need to do a workingspaceRGB to LightjetRGB conversion, and then give them that file. They would output it assuming it's just a normal file. The results you get will be better than using sRGB alone. Chris Murphy ------------------------------------------------------------------------------- From: INTERNET:Iganatz@aol.com, INTERNET:Iganatz@aol.com Date: Mon, Dec 11, 2000, 9:31 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions Whew! Thanks to all for the hearty response. A couple of points: As to Andrew's swat across my nose ("Bad dog! You never use a display profile for a Working Space. Display profiles are only used for one function, to preview images. Use a Working Space (Adobe RGB perhaps)."), I'm using the same profile as a working space in both 5.5 and 6, so if I'm pissing on the rug, at least I'm pissing on the same rug every time, and not every one in the house. Seriously, we've tried Adobe RGB and ColorMatch RGB (which was our default for a while), but found that output to Kodak Approval using the setup described in my initial post yielded proofs closest to our RGB files as previewed onscreen prior to conversion. In other words, it's what's working for us and retouchers are a superstitious lot, don't you know. We will certainly conduct more tests on wroking spaces along the upgrade path to v6, especially in light of some of the advice offered from Andrew, Peter F, Chris Murphy and the rest of the Color Theory crew, but if it ain't broke... Also, "Here's probably the issue. You don't have that Heidelberg profile loaded in your CMYK Set-Up in Photoshop 5." isn't an issue. as the profile was loaded. As to one poster's question why I'd want to ensure the same exact seps, the reason's that I was in the middle of a very large campaign and need to deliver second rounds of retouching with minor changes. A color shift, even a minor one in an image where color's been approved is a bad thing... Bottom line: A. Can't deliver the same EXACT seps out of 6 that I got from 5.5 B. 6 does seem to yield similar and in some instances superior conversions C. Once this campaign is complete, those in the studio who have not been upgraded to the new version will be, and this will no longer be an issue... Woof woof, adam sanders ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Mon, Dec 11, 2000, 10:29 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions on 12/11/00 7:29 AM, Iganatz@aol.com at Iganatz@aol.com wrote: > found that output to Kodak Approval using the setup > described in my initial post yielded proofs closest to our RGB files as > previewed onscreen prior to conversion. No surprise here. The issue is the accuracy of your display profile. The other problem is that every conversion you make is going to be different on any other machine because you are using a device specific space (your display profile) for all conversions into your output profile. Chris also illustrated why a display profile is so bad for an editing space. You can keep doing what you're doing and as long as your workflow doesn't change, fine. But the display is an every changing device. You should be calibrating and profiling it about every 300 hours or so. That means every time you create a new display profile, you're now creating a new Working Space. Now the preview and conversions can and likely will change every time you make a new display profile. If you use Adobe RGB 1998 or ColorMatch RGB, that is fixed in stone. It NEVER changes. Conversions and edits are identical today and a year from now. This is why it's such a good idea to use a Quasi-Device Independent editing space and NOT a highly device dependant space based on an ever changing product like a monitor. Andrew Rodney ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Mon, Dec 11, 2000, 11:22 AM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions > I'm using the >same profile as a working space in both 5.5 and 6, so if I'm pissing on the >rug, at least I'm pissing on the same rug every time, and not every one in >the house. Well you're not. Photoshop 5 lets you use a monitor space as a working space, but it forces it to be well behaved and gray balanced. It forces it to have a uniform gamma. Photoshop 6 doesn't do this anymore. So by using the exact same monitor setting in Photoshop 5 and 6, you can get different results when using device profiles as working spaces (or assigned profiles in Photoshop 6). So I highly recommend not doing this if for no other reason to avoid editing in an RGB space that is not gray balanced. > Seriously, we've tried Adobe RGB and ColorMatch RGB (which was our >default >for a while), but found that output to Kodak Approval using the setup >described in my initial post yielded proofs closest to our RGB files as >previewed onscreen prior to conversion. Well if the monitor isn't calibrated and profiled, this wouldn't surprise me. If it is calibrated and profiled, and you were using ColorMatch RGB and that's the only difference (CMYK Setup is the same) and you get better conversions (compared to the preview on-screen) with a monitor RGB space, that's bizarre. I've done tons of tests using sRGB and ProPhoto RGB which are DRASTICALLY dissimilar, with common real world test images, and get no visual difference for all intents and purposes. The compensation for the image's RGB color space is very good. So something else must be going on that makes your monitor space do a better job (perhaps it's lack of good gray balance and the CMYK profile's lack of good gray balance are cancelling each other out - flukes happen ALL the time). > In other words, it's what's working >for us and retouchers are a superstitious lot, don't you know. We will >certainly conduct more tests on wroking spaces along the upgrade path to v6, >especially in light of some of the advice offered from Andrew, Peter F, >Chris >Murphy and the rest of the Color Theory crew, but if it ain't broke... I know this is the inclination and I totally understand it. Most of the time I agree with it when it's crunch time. But when it's an excuse to keep doing the same thing year after year and not learn anything new about what's really going on, that's what I object to. The idea of if it ain't broke gets used so often in the graphic arts it becomes an excuse. We have layer upon layer of complexity increasing every year and it becomes more important to understand things in order to keep them simple. If you don't understand things, then you create more hacks and patches and fixes to workflow that are equally not understood, but they counteract other things you don't understand - and voila it works but is so complicated who WANTS to fix it? > Also, "Here's probably the issue. You don't have that Heidelberg profile >loaded in >your CMYK Set-Up in Photoshop 5." isn't an issue. as the profile was loaded. For separations it should be set with black point compensation ON and rendering intent set to either perceptual or relative colorimeteric (image specific). For previewing CMYK images you need to change it to black point compensation OFF, and rendering intent to relative colorimetric. That makes a BIG differnce in preview. Photoshop 6 does this for you automatically. > As to one poster's question why I'd want to ensure the same exact seps, >the reason's that I was in the middle of a very large campaign and need to >deliver second rounds of retouching with minor changes. A color shift, even >a minor one in an image where color's been approved is a bad thing... Yeah in that case it's better to keep using the same version of Photoshop until that project is finished and not change Photoshops in the middle of a project. Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Mon, Dec 11, 2000, 12:15 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions >You should be calibrating >and profiling it about every 300 hours or so. At least. With an average work day that's only once every 42 days. Maybe you meant every 30 hours Andrew? I think this is very monitor and application specific. Some people can go a month or two, and others need to do it every morning after their monitor warms up for an hour. And some people need a Barco which does it something like 10 times a second. > That means every time you >create a new display profile, you're now creating a new Working Space. That's a *really* good point. That is a MARVELOUS point. I need to file that back in my head. I've thought of this before but not in this thread. That's a huge big deal for three reasons: If you want images to look the same every time you look at them: 1.) When the working space changes because your monitor profile is now different, if you embed profiles in images, it means you will have to *convert on open* every single legacy file. Even one from a few weeks ago. You will have to convert it from the old space to the new one. 2.) #1 applies for each monitor in a workflow. Each workstation has a separate working space!!! Now Photoshop 6 allows you to work in someone else's working space automatically if you set it up that way so you wouldn't have to convert. But still it means each workstation and image will be using different editing spaces. One minute you're working with your monitor's lack of gray balance (needing more blue) and on another image you're working on someone elses image and they have too little blue. Gray balance is a MESS in this workflow! 3.) If you *don't* embed profiles, then you have no idea how that image looked a week ago. You will have to assume a new and different working space (your current monitor profile) which means each time you reprofile the monitor, your images will look different! So this is a dead end workflow especially with Photoshop 6 that does not force display profiles to be well behaved like previous versions of Photoshop. Get rid of those display profiles used as working spaces and embedded into images and STANDARDIZE on a well behaved color space. > If you use Adobe RGB 1998 or ColorMatch RGB, that is >fixed in stone. It NEVER changes. Either on the same workstation or on different workstations, even if they are using Photoshop 5. Otherwise in Photoshop 5 they either MUST convert, or if they don't, they won't see it the same way you did. They would have to have Photoshop 6 to avoid this problem (to some extent). Chris Murphy ------------------------------------------------------------------------------- From: Andrew Rodney, INTERNET:andrew@digitaldog.net Date: Mon, Dec 11, 2000, 12:24 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions on 12/11/00 10:14 AM, Chris Murphy at lists@colorremedies.com wrote: > At least. With an average work day that's only once every 42 days. Maybe > you meant every 30 hours Andrew? I took the 300 hours from Barco. That's what they recommend. But Barco's are so darn accurate and self adjusting that it's probably a good idea to calibrate and profile a non Barco MUCH more often. Andrew ------------------------------------------------------------------------------- From: "shAf", INTERNET:michael@shaffer.net Date: Mon, Dec 11, 2000, 12:55 PM RE: RE: [ColorTheory] 6.0 Profile to Profile conversions Chris writes ... > ... > > So this is a dead end workflow especially with PS6 > that does not force display profiles to be well > behaved like previous versions of Photoshop. > ... Please point us at a reference which explains these PS6 differences with respect to PS5. All we have been led to believe so far, is that the only difference for "DUMC" is that is always on in PS6. shAf:o) ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Mon, Dec 11, 2000, 1:36 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions >I took the 300 hours from Barco. That's what they recommend. But Barco's are >so darn accurate and self adjusting that it's probably a good idea to >calibrate and profile a non Barco MUCH more often. Ahh OK that makes sense. Yeah most monitors aren't self adjusting so for most color critical situations once a week is enough is what I tell my customers. For highly color critical environments, daily after 1 hour warm-up is not unreasonable. For those who think this is overkill, a brand new monitor can be successfully calibrated to 5000K for only the first three months of its life. After than it's questionable. That's how rapidly the phosphors decay. They are like Bic lighters that cost 3000 times as much. Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Mon, Dec 11, 2000, 1:34 PM RE: RE: [ColorTheory] 6.0 Profile to Profile conversions >Please point us at a reference which explains these PS6 differences >with respect to PS5. All we have been led to believe so far, is that >the only difference for "DUMC" is that is always on in PS6. I don't know. Andrew? The primary difference in regards to RGB support in Photoshop 6 is that it now lets you use *any* RGB profile for a document space (or working space). This means an RGB inkjet profile can be assigned to a document (or as the RGB working space). In Photoshop 5 this wasn't the case. The only kind of device profile you could use as the RGB Working space in Photoshop 5 (and earlier for that matter) were matrix type monitor profiles. The matrix monitor profile has little information in it. The red primary, the green primary, and the blue primary (a triangle), the white point, and the gamma for each channel. Photoshop 5 and earlier doesn't use the gamma for each channel. It uses only one (I don't know if it averages them or just picks one) which makes the space more uniform. It also forces R=G=B to be neutral gray, even if the profile says the monitor behavior isn't gray (which usually it isn't). So this is a concern with Photoshop 6 because unless you use a mathematical profile like ColorMatch or Adobe RGB, who knows what gray is in RGB? Chris Murphy ------------------------------------------------------------------------------- From: Bob Smith, INTERNET:rmsmith@calpha.com Date: Mon, Dec 11, 2000, 1:47 PM RE: Re: [ColorTheory] 6.0 Profile to Profile conversions Andrew Rodney wrote: > If you use Adobe RGB 1998 or ColorMatch RGB, that is > fixed in stone. It NEVER changes. Conversions and edits are identical today > and a year from now. This is why it's such a good idea to use a Quasi-Device > Independent editing space and NOT a highly device dependant space based on > an ever changing product like a monitor. a personal testimonial on the value of standard working spaces and embedded profiles: I'm in the process of working on a catalog for a client that I have been shooting for since my earliest days of doing any sort of digital work (93). This is a relatively small client for whom I've done maybe two shoots a year. The catalog we're currently working on is a major undertaking for these people involving more color photography than they've ever used before. The products are custom made pieces of furniture so its not always easy to get samples to photograph. We're trying to use as many existing images as possible. I recently dug through the archives and retrieved every image I've ever digitized for these people. The images processed since Photoshop 5 were simple as can be to deal with. I had experimented with working spaces for a while and as a result files for these people were in a variety of RGB spaces depending on what I happened to be using at the time. However since every file was tagged and a profile embedded it was simple to convert to my current standard, Adobe RGB. All of those images looked fine on screen and matched each other fairly well. The files from the days of Photoshop 2.5 through 4 were another story. Many, but not all of those images were done using a hardware calibrated display. However there was still a substantial variation in color balance and tone to the images and no embedded profile for reference. I couldn't just "assume" ColorMatch RGB and be in the ballpark. It has taken quite a bit of manual tweaking to get those old images to look like the ones archived from Photoshop 5 forward. Bob Smith ------------------------------------------------------------------------------- From: "shAf", INTERNET:michael@shaffer.net Date: Mon, Dec 11, 2000, 3:49 PM RE: RE: [ColorTheory] 6.0 Profile to Profile conversions Andrew writes ... > > on 12/11/00 11:33 AM, Chris Murphy at lists@colorremedies.com wrote: > > > So this is a concern with Photoshop 6 because unless > > you use a mathematical profile like ColorMatch or > > Adobe RGB, who knows what gray is in RGB? > > This is easy to test and Illustrate what Chris is talking about. > ... Thanx Chris & Andrew ... and if I might, as long as we're on the subject of profile conversions and lack of documentation for PS6, can I ask ... where I can find better descriptions of the new options for the CMM rendering engines used. I wonder if they don't default to common use in 8bit space, while most of my conversions are with 16bits/channel (photographic, of course). shAf:o) ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Mon, Dec 11, 2000, 4:43 PM RE: RE: [ColorTheory] 6.0 Profile to Profile conversions > Thanx Chris & Andrew ... and if I might, as long as we're on the >subject of profile conversions and lack of documentation for PS6, can >I ask ... where I can find better descriptions of the new options for >the CMM rendering engines used. There really aren't any new options that I'm aware of. > I wonder if they don't default to >common use in 8bit space, while most of my conversions are with >16bits/channel (photographic, of course). ColorSync CMM conversions (even 8-bit/channel images) are usually converted with 16-bit precision. It's CMM dependent. Apple's and Kodak's go through 16-bits/channel precision. If you use Adobe ACE engine, then all transforms are 20-bits/channel I think. It's more than 16-bits/channel in any event. Chris Murphy ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Wed, Dec 13, 2000, 12:31 PM RE: RE: [ColorTheory] 6.0 Profile to Profile conversions > If you choose the "advanced" setup for the color preferences ... you >get to realize several new rendering "intentions", what's more, the >defaults are not the same as PS5. intentions? I don't have such a thing. The Intents are the same as in Photoshop 5, although you only got to use them if you were using an ICC profile; since you can only use ICC profiles in Photoshop 6, an intent is mandatory. The defaults for pretty much everything are the same in Photoshop 6 as for Photoshop 5 as well. (There are new presets which can make the default behavior different, but Black Point Compensation has always been on in Photoshop since 3 or 4, and the rendering intent is relative colorimetric since Photoshop 2.5.) > I realize the computations are done with high precision, but (for >example) the "dithering" checkbox I believe is for resulting 8bit >quantization errors. Yes. That's a new feature. > Two of the rendering "intentions" would seem >appropriate for photographic images, but I've found only two small >paragraphs which mention them or their application. I don't have such an option called intentions. Chris Murphy ------------------------------------------------------------------------------- From: "shAf", INTERNET:michael@shaffer.net Date: Fri, Dec 15, 2000, 3:28 PM RE: RE: [ColorTheory] 6.0 Profile to Profile conversions Chris writes ... regarding: > > If you choose the "advanced" setup for the color > > preferences ... you get to realize several new > > rendering "intentions", what's more, the > > defaults are not the same as PS5. > > intentions? I don't have such a thing. The Intents are the > same as in Photoshop 5, although you only got > to use them if you were using an ICC > profile; since you can only use ICC profiles > in Photoshop 6, an intent is mandatory. Sorry, I was referring to intents ... and after examining a PS5 install vs a new PS6, ... yes, the intents offered are the same (but, see below). > > The defaults for pretty much everything are the same in > Photoshop 6 as for Photoshop 5 as well. > ... The engines offered for my PS5 workstation are "built-in" (the default), and "Kodak ...", and the "default" intent is "perceptual (images)". however ... The engines offered for PS6 are "Adobe ACE" (assumed to the equivelent to "built-in"), and "Microsoft ICM" (which is new). (It would be interesting to know what the alternative to Adobe ACE is on a MAC installation.) The intents are the same, but my PS6 installation defaulted to "relative colorimetric" and there is no implication that "perceptual" is for images(?) On my Windows workstations, these apparent changes confused me a little ... and there seems to be little information available regarding why. Your implication that nothing has changed might imply PS6/Mac doesn't refer to "Microsoft" at all, and that I should still be using the perceptual intent(?) shAf:o) ------------------------------------------------------------------------------- From: Chris Murphy, INTERNET:lists@colorremedies.com Date: Fri, Dec 15, 2000, 5:01 PM RE: RE: [ColorTheory] 6.0 Profile to Profile conversions > The engines offered for PS6 are "Adobe ACE" (assumed to the >equivelent to "built-in"), and "Microsoft ICM" (which is new). Adobe (ACE) is essentially built-in. I have no idea what MicroSoft ICM is. Perhaps this means that Photoshop 6 now uses ICM on Windows. >The intents are the same, but my PS6 installation >defaulted to "relative colorimetric" and there is no implication that >"perceptual" is for images(?) Well Photoshop 5 was unique in that it had two different ways you could separate color. The CMYK setup would let you use the classic engine which was the "built-in" radio button where you select ink colors, dot gain, total ink, GCR/UCR, black generation, black ink limit, etc. That part in Photoshop 5 and earlier used relative colorimetric. The part in Photoshop 5 that used ICC profile defaulted to Perceptual which in my opinion was wrong but that's what they did. I also found it really misleading for them to put "Perceptual (Images)" implying that the perceptual intent was for images and the other intents weren't. I don't know what the breakdown is, but I feel most images should be converted relative colorimetric. Only those images that have bright vivid saturated colors that cannot be reproduced in the destination's gamut should use perceptual. And in some cases you'll find it's still OK to use relative colorimetric just like we've been doing for years in previous versions of Photoshop. The perceptual intent reduces the saturation of the image in order to make room in the destination gamut for out of gamut colors - so all of them can be represented. Represented doesn't mean exactly reproduced. For example, a hibiscus. A nice very very satuated red flower that has lots of detail in the petals. Nearly of those different colors of red are out of gamut for the vast majority of devices. Using relative colorimetric rendering, you will likely see posterization because different colors of red get mapped to the gamut boundry - i.e. the same color of red. So you lose detail. Using perceptual, everything is desaturated a bit so that more of the different colors can be represented in print. The result is a slightly desaturated image, but the hibiscus will output with good detail. > Your implication that nothing has changed might imply >PS6/Mac doesn't refer to "Microsoft" at all, and that I should still >be using the perceptual intent(?) No. I thought you meant that you actually had different rendering intents vs. having different defaults. Chris Murphy