Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: Roscolo on January 06, 2012, 03:08:11 pm
-
I am doing more and more fine art repro (paintings, oils, watercolors). Printing on HP z3100. NEC P221W monitor / Spectraview II calibrator.
My workflow has been primarily making adjustments on screen, which works fairly well, as depending on the painting I have to make changes to selective areas anyway. Some paintings take 5 minutes to get an acceptable proof. Some can take a few hours.
However, with the level of work I'm now getting I need to speed things up a bit if I can. Would like to hear from anyone using the Colorchecker Passport with fine art reproduction. Is it accurate enough / helpful for this kind of work? Any other product you would recommend?
thx
-
It works great! I have started using it with my wife's pastel drawings. It has taken hours off the process. Can't recommend it enough. Even shooting in direct sunlight the reproductions are excellent with almost no tweaking. None of the canned Lightroom or ACR profiles can even come close. It is especially effective with difficult colors like saturated reds and deep blues, which were previously nightmares to get right. I shoot a chart for each piece. On the 30 x 44 originals it's enough to have somebody just hold the chart near the center of the print, but be careful to position their arm not to get too much bounce off the flesh.
In the old days we used to shoot "Lilly" charts down at the corner of the Ektachromes, but it always seemed like the charts had a color balance a little different than the piece, which I believe had to do with surface texture and sub-surface scattering. The Passport system doesn't seem to have that problem, bless its chromatic little soul.
-
It is especially effective with difficult colors like saturated reds and deep blues, which were previously nightmares to get right.
I believe you have read my mind. Thanks for the info, Bill. Your recommendation is good enough for me!
One more question: I just built a new workstation, but I would like to be able to put it on my old one as well. Can you install the included software on more than one computer?
-
Randy Hufford has a great video on the subject of photographing fine art and reproducing it in print. If you are into fine art reproduction you might enjoy what he has to say. His laid back Hawaiian style throws some, but the man knows his stuff.
Ron
-
One more question: I just built a new workstation, but I would like to be able to put it on my old one as well. Can you install the included software on more than one computer?
Yes, I believe you can download the software for free from X-Rite and use it on as many computers as you want, once you own the color checker itself.
-
Well, bought the Passport Colorchecker and just tried it out. Picked a painting with a lot of purples / magentas. Shot the chart, shot artwork under exact same conditions, built profile, applied profile, applied white balance.
On the training video the lady on the videos builds a profile and then applies it and says, "when you apply the profile you'll see THOSE COLORS ARE JUST POPPING RIGHT OUT!" Then she goes on to say, "You'll see it gives you a much richer, much more elegant look to your images." lol...what the hell is she talking about?!? The purpose of color management is to get accurate repro of colors in a scene, not to "make colors pop out" or appear "richer, more elegant." Good luck communicating color management using those terms.
So, I was a little harsh on my first review, so I've edited this. I've played with with the Colorchecker Passport. It's OK. Maybe a slight advantage, but I can't say it's any huge advantage over Adobe's presets for repro of artwork.
-
What? I think the Passport it's the cat's meow for art repro! Something's probably wrong.
How's the exposure? If you have a lot of LR adjustments applied the accuracy of camera profiles is compromised. What kind of lighting are you using? Are any of the color channels clipped on your original? Your camera should be set to Adobe RGB or other wide gamut color space. Etc.
In my case I usually shoot art in the afternoon sun with the sun about 45 degrees to the art, no nearby strongly colored objects such as a brick walkways, cars, etc. I carefully check that I'm not clipping any of the RGB histograms, adjusting the exposure as needed, that's a little easier if the art is against a gray matboard or such so as not to distort the histograms with extremely bright or dark areas from the non-art background. If you're just looking at the composite "luminance" histogram it's easy to clip one or more colors without realizing it, especially red. If I take a new profile for the particular photograph I usually only have to adjust exposure, brightness and contrast, and only very minor tweaks to color and saturation.
Don't give up! It really works.
-
The painting I have chosen is watercolor landscape with lots of hues of magentas and blues. I shoot with two 3200K lamps on either side of the artwork in a blacked out studio, as I have for years. Shot the painting. Then shot the colorchecker. Built my profile per instructions. Did it again just to be sure. Then white balanced and saved as an Adobe Raw Preset. Applied to painting. Everything was too magenta. So I built another preset and white balanced using the white balance squares for color adjustment on the Colorchecker, essentially making the next preset I built "cooler." But, nay, now a little too cool. The blues are overtaking the magentas. So, I'm getting close, but already at this point the passport colorchecker offers no real savings in accuracy or time over just using the Adobe raw present for Camera Neutral or sometimes Camera Standard. I usually get them right after this many proofs anyway.
I'm probably going to return this. Honestly, if it was priced where it should be, I don't know, $20-$30, I might keep it for general use on portraits or something. Not sure if it would offer any big advantage there over Adobe's presets either. For painting repro, $100 for ballpark color accuracy is too much...I can do ballpark color with my eyes as fast as this.
Glad it works for you. If it doesn't help me match color in 3-4 tries or less, it's useless for me because I can usually get my colors right in 3-4 tries just using my eyes. I'll give it a try on a less challenging painting, but honestly, it's the challenging paintings I bought it for.
-
Hi,
In art repro you would try tio achieve "scene referred rendering". Normal color rendition is intended to be more pleasant than faithful.
I down't know the setting in LR and Camera Raw but parking each control at 0 is a good idea.
Best regards
Erik
The painting I have chosen is watercolor landscape with lots of hues of magentas and blues. I shoot with two 3200K lamps on either side of the artwork in a blacked out studio, as I have for years. Shot the painting. Then shot the colorchecker. Built my profile per instructions. Did it again just to be sure. Then white balanced and saved as an Adobe Raw Preset. Applied to painting. Everything was too magenta. So I built another preset and white balanced using the white balance squares for color adjustment on the Colorchecker, essentially making the next preset I built "cooler." But, nay, now a little too cool. The blues are overtaking the magentas. So, I'm getting close, but already at this point the passport colorchecker offers no real savings in accuracy or time over just using the Adobe raw present for Camera Neutral or sometimes Camera Standard. I usually get them right after this many proofs anyway.
I'm probably going to return this. Honestly, if it was priced where it should be, I don't know, $20-$30, I might keep it for general use on portraits or something. Not sure if it would offer any big advantage there over Adobe's presets either. For painting repro, $100 for ballpark color accuracy is too much...I can do ballpark color with my eyes as fast as this.
Glad it works for you. If it doesn't help me match color in 3-4 tries or less, it's useless for me because I can usually get my colors right in 3-4 tries just using my eyes. I'll give it a try on a less challenging painting, but honestly, it's the challenging paintings I bought it for.
-
The purpose of color management is to get accurate repro of colors in a scene, not to "make colors pop out" or appear "richer, more elegant."
Not really. See:http://www.color.org/ICC_white_paper_20_Digital_photography_color_management_basics.pdf
-
Not really. See:http://www.color.org/ICC_white_paper_20_Digital_photography_color_management_basics.pdf
Yes, really. For me and my customers color management for repro of artwork has everything to do with accurate repro of colors. I assure you, my customers want the colors in their prints to match the colors in their artwork. Thanks for the info. but the 2005 document you linked to says nothing about accurate color reproduction of works of fine art - the subject of this thread.
That said, I did another print from a more representative "average" painting and got a much better result from the Colorchecker Preset I created. It was pretty close, I made a couple of tweaks, lightened the print a bit, removed some saturation, and had a good matched print.
I can say this - the Colorchecker seems to add too much saturation as a default - seems first thing I need to knock about 8-10 pts. off saturation from what the Colorchecker profile gives. And actually, when you watch the crazy "training video" that emphasizes making "colors pop out" and "richer, more elegant" results, you can see when she applies the profile even on the video the flowers are oversaturated when the profile is applied, hence the "pop out" effect. Good for snapshots of granny's flowers, not good for much else. Not sure if that makes granny's flowers "elegant" - lol. Now that I'm learning some of the weaknesses of the Colorchecker profile, I can edit these out of the presets I create using it. Primarily that seems to be a default setting of the Colorchecker software to oversaturate.
So, useful tool. Overpriced. Software needs some work. I'll just edit my presets. I have several paintings to do today, so I should know by the weekend if it warrants keeping or returning. But the second painting I achieved much better results and faster as that painting represents more of an average scene, not so many overlapped hues of similar color, and admittedly challenging colors to reproduce, particularly on the z3100 - magentas, purples - like the first painting. The first painting I tried was perhaps too challenging to expect too much from the Colorchecker.
-
For me and my customers color management for repro of artwork has everything to do with accurate repro of colors.
And based on your understanding of the ICC white paper, that would be accurate how? Colorimetrically and defined how numerically and to what dE? Since you are working with output referred color, this has to be a subjective visual match. Otherwise, what metric would you use to define this being accurate?
I assure you, my customers want the colors in their prints to match the colors in their artwork.
You mean they tell you they believe the two visually match?
Thanks for the info. but the 2005 document you linked to says nothing about accurate color reproduction of works of fine art - the subject of this thread.
Because there are two ways to define accurate. One is subjective so anyone can disagree. The other is by measuring the colors in the original which absolutely will not numerically match the rendering you have to show to a customer.
That said, I did another print from a more representative "average" painting and got a much better result from the Colorchecker Preset I created. It was pretty close,
You define ‘pretty close’ how? To what degree of accuracy metric?
So, useful tool. Overpriced. Software needs some work.
If it gets you closer to your subjective idea of matching, then you have to decide if it is worth the price. If you assume using it will either produce accurate colorimetric match (it will not) or a prefect subjective visual match, then no, it isn’t going to and I suspect isn’t marketed to do this.
-
Accurate repro of colors in fine art, for my customers, means when they bring me a watercolor painting, and I create a print of that painting, that one can placed the print on top of the painting and the colors in the print match the colors in the painting.
No, the colorchecker passport is probably not meant specifically for this type of work. That's why I started this thread. All one has to do to see what the colorchecker passport is marketed towards is to watch the company's own training video. The company rep states it's to make colors "pop out" and be "richer...more elegant." Can't get much more subjective than that! :) In all honesty had I watched the video I probably would not have purchased. It's not a good way to demonstrate how accurate your color management tool is when the video used to demo the tool uses granny's flowers as the demo subject. The color could be off by a mile and granny's flowers would still "pop out." :)
So, as I stated, I've identified some faults for my current use, but even in my commercial work, I don't consider adding saturation to an image to make it "pop out" desirable or even tolerable. Great for granny's flowers. Not necessarily good for mixed lighting architectural interiors. Certainly not a default I want to apply to every image. Like I said, it's easy enough to edit those drawbacks out. For repro of fine art, so far, I'm looking at removing 8-10 pts. of saturation from the colorchecker profile.
I'm going to continue to go through some more paintings. Right now I stand by that this tool, while useful, considering the seller recommends buying one every two years, should probably carry a price tag of no more than $40. Definitely stand by that the software needs some work, but realistically anyone using this is going to have to edit and create their own presets anyway, so I suppose it falls under the heading of "good enough." The company must feel the same way as they haven't updated the software in over 18 months.
The first painting was a bear by any standard. I'm going to give it a full run through with the work I have to complete, so I will withhold final judgment on whether to return it or not until I've done several more paintings. Now that I've identified the saturation problem, my workflow is speeding up. I also have a commercial architectural shoot at a hospital on Monday, so I'm going to give the colorchecker a run through there in some mixed lighting situations.
-
Accurate repro of colors in fine art, for my customers, means when they bring me a watercolor painting, and I create a print of that painting, that one can placed the print on top of the painting and the colors in the print match the colors in the painting.
OK, so there is no subjectivity here? When you are ready to present the repro to your client, they always agree there is a match? You and all your clients have the same (or ideal) color perception?
IF we agree there is some subjectivity here, the term accurate becomes a problematic term to use and apply to a profile.
If the idea is, you shoot a ColorChecker with a mere 24 colors and some profile is magically built that always produces a visual match of your original, the idea is off base. I wish it were that easy; point the camera at some artwork, build a profile, a prefect match. We both know it is far more difficult. We both know the pigments, dyes and other colorants in an original and that off your printer can be vastly different. That we can and do see metameric failure. That some colors selectively need to be edited. FWIW, this was no different with film!
Once again, if the new profile gets you closer to your subjective rendering of the original, then maybe it is worthwhile at the price point. But only you can make that call. But believing it is going to produce some prefect match isn’t going to happen. And color management doesn’t in any way ensure this. Color management isn’t push button “correct” color. What color management does is give numbers (the only thing computers understand) a meaning but those numbers could be totally inappropriate to produce the color you desire. That is why you have to massage those numbers.
-
When you are ready to present the repro to your client, they always agree there is a match? You and all your clients have the same (or ideal) color perception?
Once again, if the new profile gets you closer to your subjective rendering of the original, then maybe it is worthwhile at the price point. But only you can make that call.
To your first question, the answer is yes, when my customer and I are standing in front of the painting and the print, we see the same thing. I've never had a situation arise where someone told me a specific area did not match and I couldn't see what they were talking about. About the only time we don't agree is when I try to talk them into a color match that we all can see...doesn't match. :)
As to the second point, I think I can safely call it now. The colorchecker is not worthwhile for doing the type of work I am doing. Part of this may be because I've been doing this for years now, and I have a good eye, and some pretty good presets I've saved over time, and one of those presets usually get me close to an accurate result. Would the colorchecker be more useful if I didn't already have those good presets I've created? Probably. My lighting set-up is also very consistent, so it could prove useful for someone having to photograph artwork under less than optimal varying light conditions. But right now my advice to someone producing prints of paintings who is shooting in a studio under consistent conditions and has some experience reproducing paintings is the colorchecker is not going to save you much if any time printing reproductions of fine art (paintings).
-
I would like to add this to the record for future generations of art reproduction mavens using the X-rite Passport device with LightRoom.
With the Passport, after making the profile and re-starting LightRoom you need to first load the new profile, and then ALSO use the eye-dropper tool in the LightRoom White Balance control panel to sample one of the grey patches on the Passport image. Only then will your image start to look its profiled best. I like to use the third grey patch down underneath the letters "ecker." Then go on and adjust exposure, brightness, contrast, and then go on to color tweaks. Incidentally the black textured case of the Passport is very nice for getting the shadow areas and overall contrast right.
Also, if you take the WB sample from a noisy image you need to carefully look at the pixel-level magnification that LR presents during the sample taking process. Make sure you do not sample an area with color noise distortion. After taking the sample, move the eye-dropper around on the grey patch while watching the RGB values to make sure you are pretty neutral across the patch, and re-sample the WB if you see an overall imbalance in the RGB numbers.
As a further proof of WB quality (and indirectly of profile quality), you can check the RGB values for the various density grey patches to make sure the RGB numbers are mostly the same, once again moving the pointer around to average out the influence of noisy pixels.
Note that the 2 columns of 5 not-quite-grey patches on the panel opposite the color patches are designed to warm or cool the overall image, and should NOT be used to set a neutral WB.
Hope this is helpful. I just love that little Passport. But I do so wish I didn't have to restart LR to apply the new profile, and I wish the profile loading screen was a little bigger and the overall LR file managing for profiles was a little slicker.
And if you have clipped color channels all bets are off, so be watching all three histograms when you hit that shoot button.
-
So, I was a little harsh on my first review, so I've edited this. I've played with with the Colorchecker Passport. It's OK. Maybe a slight advantage, but I can't say it's any huge advantage over Adobe's presets for repro of artwork.
That was my experience as well. I found the PS RAW profile for my camera (GH2) worked slightly better than the Colorchecker profile built with their software or one built with the Adobe DNG editor and none were really right on. I finally wound up using the PS RAW profile and some tweaks and can get a very decent result which isn't too bad to get within acceptable limits.
-
I do fine art reproduction for a living and have found that the Passport profiles are a little too vivid, plus the 24 samples are not very evenly distributed in the color spectrum, so there are some pretty big holes that the software has to guess at (I know MacBeth's motivation for choosing the colors they chose, but it doesn't help me much). I've had better luck with the Qp card, and their free Qpcalibration software. The 35 colors in the Qp chart are more helpful for art reproduction purposes, at least for me, and I find that I start with a closer approximation using that system than what I was using with a MacBeth chart and DNGProfiler (which I liked better than the Passport software). Photo printing is so much easier to do than art reproduction, just because the standards are so high when side by side comparisons are as easy as they are with art.
-
I do fine art reproduction for a living and have found that the Passport profiles are a little too vivid, plus the 24 samples are not very evenly distributed in the color spectrum, so there are some pretty big holes that the software has to guess at (I know MacBeth's motivation for choosing the colors they chose, but it doesn't help me much). I've had better luck with the Qp card, and their free Qpcalibration software. The 35 colors in the Qp chart are more helpful for art reproduction purposes, at least for me, and I find that I start with a closer approximation using that system than what I was using with a MacBeth chart and DNGProfiler (which I liked better than the Passport software). Photo printing is so much easier to do than art reproduction, just because the standards are so high when side by side comparisons are as easy as they are with art.
Wow. Didn't know about QP. That looks better and the price is more in line with what I think something like this is actually worth. Self-adhesive as well. Perfect. Well, that's what I love about the LL forums. Thank you.
Interestingly enough, looking at the QP website, it looks like they will be releasing software that will build profiles using the Passport Colorchecker, which is what I immediately thought of when I saw how saturated their profile was: that maybe someone had a private, donation-based, or even open source software project to make a better profiler using the Colorchecker chart. Nice to see someone else sees the need (and an opportunity.)
Thank you so much for this information.
-
I'm not sure about the self-adhesive part. I bought the chart that is roughly the same size as the MacBeth Color Checker. It's not as beefy as the classic Color Checker, but works well enough. I still use my Color Checker for gray balance and adjusting the tonal range when I import into Lightroom. I have it set with a preset to use my Qp chart profile on import, then I adjust the gray balance and tonal values of the six neutral squares from the Color Checker that have published values to match to. For some reason, the gray squares of the Qp are a little too warm for gray balance, and are different than the three square strips that Qp sells which are very neutral. I bought my chart direct from the manufacturer, and got it in about two weeks time from Europe to Hawaii.
PS: I also use a Z3100.
-
Thank you, thank you, thank you. I just saw the chart you are referring to on their site. Ron, can I ask you where you bought yours? All the pricing on their site is in Euros.
EDIT: Never mind. I see that you bought it from Europe. Sounds much more appropriate for the work I'm doing, so I'll do some digging and order from there if I can't find a stateside source.
-
They normally cary it at B&H, but they were out of stock when I ordered mine. As best I recall, it was slightly cheaper ordering it from the manufacturer, also, because B&H charges a premium to ship to Hawaii, and Qp sends it via mail with no surcharge.
-
I work as a consultant in fine art reproduction, they goal for my customers has always been to reproduce colours exactly as they appear in the original painting, I can offer the following advice:-
1. Forgot Passport it does not have enough colours to be useful for this application, I use a chart with a palette of over 1000 patches that are tuned to the gamut of art medium being photographed ie oil, water colour etc. If producing and measuring your own chart is out of the question use a colorchecker SG or Kodak IT8 not ideal by any means but a step up from Passport.
2. Shoot RAW but do not use ACR for RAW conversion it does not have a wide enough gamut to accommodate all the colours you may need to preserve, it is also built for speed not accuracy detail smoothing and demosaic errors will appear. I use a heavily modified version of Rawtherapee that includes slow but accurate demosaicing my own input and output spaces tuned to fine art and is floating point for greater accuracy and detail.
3. Use a high quality balanced light source around D50, and set white-balance in the camera, you should not have to play with the white balance after the image is taken.
4. A good LUT based input ICC profile is critical to reproduce colours such as Cobalt Blue accurately.
I hope this is of some help!
-
2. Shoot RAW but do not use ACR for RAW conversion it does not have a wide enough gamut to accommodate all the colours you may need to preserve
Could you elaborate on this?
-
3. Use a high quality balanced light source around D50, and set white-balance in the camera, you should not have to play with the white balance after the image is taken.
Set the camera's white balance where?
I paint under a combination of fluorescent lights that give off D50. At least the color band - or no band - checker from Bruce Fraser's Real World Color Management doesn't show bands... but I photograph with my camera set at Auto White Balance, not D50.
I've tried various other settings, Standard, Portrait, Landscape, etc., and believe Neutral gives me best results. Anyone's thoughts?
...
Is this 30 color card the B&H QP card in question?
http://www.bhphotovideo.com/c/product/286652-REG/QP_Card_GQP201.html (http://www.bhphotovideo.com/c/product/286652-REG/QP_Card_GQP201.html)
...
I once had a giclee made on Epson 'watercolor' paper that was such a close match to my original watercolor that with both images matted, most people could tell which was the print and which was the original. I could tell, but I had to look closely. It wasn't color so much as I knew the difference in the texture of the papers. I believe that's exactly what artists who make limited edition prints want.
Paz
-
Is this 30 color card the B&H QP card in question?
http://www.bhphotovideo.com/c/product/286652-REG/QP_Card_GQP201.html (http://www.bhphotovideo.com/c/product/286652-REG/QP_Card_GQP201.html)
This is their current offering for the larger size card that I think the profiling software is built for. It doesn't look like B&H carries this card anymore.
http://www.qpcard.com/en_b2c/products-on-index/qpcard-203-book-colorcard-profile.html
-
Take a custom white balance with your camera, this will set the optimum white balance for your light source.
Set the camera's white balance where?
I paint under a combination of fluorescent lights that give off D50. At least the color band - or no band - checker from Bruce Fraser's Real World Color Management doesn't show bands... but I photograph with my camera set at Auto White Balance, not D50.
I've tried various other settings, Standard, Portrait, Landscape, etc., and believe Neutral gives me best results. Anyone's thoughts?
...
Is this 30 color card the B&H QP card in question?
http://www.bhphotovideo.com/c/product/286652-REG/QP_Card_GQP201.html (http://www.bhphotovideo.com/c/product/286652-REG/QP_Card_GQP201.html)
...
I once had a giclee made on Epson 'watercolor' paper that was such a close match to my original watercolor that with both images matted, most people could tell which was the print and which was the original. I could tell, but I had to look closely. It wasn't color so much as I knew the difference in the texture of the papers. I believe that's exactly what artists who make limited edition prints want.
Paz
-
2. Shoot RAW but do not use ACR for RAW conversion it does not have a wide enough gamut to accommodate all the colours you may need to preserve,
ACR doesn't have a "gamut". It uses a working space which is defined by the user and the working space is what defines the gamut that it is capable of representing when the final image is output. ACR is capable of using ProPhoto RGB at 16 bits. If this can't accommodate "all the colors you may need to preserve", ACR isn't what I'd be worried about, it's your choice of printers because I'm not aware of any printer than can reproduce all of ProPhoto RGB.
It certainly would be nice if ACR and Lightroom supported ICC profiles for various cameras but they don't. Capture One does.
Cheers, Joe
-
It would be really helpful if HueLight posted some 100% crop comparisons to show what's not captured in ACR over what can be pulled out of the scene gamut using his Rawtherapee method.
From my experience getting dead to nuts accurate scene captures on a wide range of objects made of acrylic to dyed wool processing with ACR, the quality and amount of light used to light the subject does most of the heavy lifting in getting the most accurate reproduction shooting Raw with a DSLR.
-
Hi,
My impression is that ICC profiles are not really appropriate for capture devices. I may be wrong on this, of course.
Best regards
Erik
ACR doesn't have a "gamut". It uses a working space which is defined by the user and the working space is what defines the gamut that it is capable of representing when the final image is output. ACR is capable of using ProPhoto RGB at 16 bits. If this can't accommodate "all the colors you may need to preserve", ACR isn't what I'd be worried about, it's your choice of printers because I'm not aware of any printer than can reproduce all of ProPhoto RGB.
It certainly would be nice if ACR and Lightroom supported ICC profiles for various cameras but they don't. Capture One does.
Cheers, Joe
-
Huelight-
Could you be more specific about which 1000+ patch color chart you use?
-
Hi,
My impression is that ICC profiles are not really appropriate for capture devices. I may be wrong on this, of course.
Best regards
Erik
I think they're quite appropriate. The question is, are they practical? For a controlled environment with controlled lighting that is the same for every single shot... sure. Capture One uses ICC profiles for every camera they support. Exactly how they build those profiles, and how they tweak them to work across a range of different lighting conditions (not only different color temperatures but different spectral power distributions) I'm not entirely sure. They weren't always great though. Back when I shot Canon I had a 1DMkII. I was using Capture 1 at the time and the Phase One provided profiles were not very good at all. The Magne Etc. profiles however were fantastic. Unfortunately he stopped developing them for newer cameras due to piracy issues, it just wasn't profitable for him anymore.
The same applies to DNG profiles, they're not terribly useful for lighting conditions, other than those that you shot the target in. Fortunately the Passport Color Checker target is small enough that it fits in a shirt pocket and can be whipped out for a quick shot if you think the lighting might pose some challenges.
Cheers, Joe
-
The really big difference between an ICC profile and a DNG profile is one is output referred and the other is scene referred. Now if you don’t think all the rendering in a raw converter plays much a role in the entire process of profiling, then an ICC profile is fine and as shewhorn points out, that means treating the camera a lot like a scanner. Lock stuff down in a very tight shooting condition.
The other issue is that when you are dealing with an output referred profile process, the target plays a big role. You can’t build an ICC camera profile who’s gamut is larger than the gamut of the target used to build said profile. They are a bitch to shoot properly too. Look at the original ColorChecker DC target, created when GMB tried to provide camera profiles. Lots of colors but many on a very glossy material to extend the gamut. Only issue was, a tiny amount of flar or reflection off the target and a butt ugly profile. The next target they built, no shiny patches. Easier to shoot, lower gamut profiles. They finally gave up on the process.
-
ACR doesn't have a "gamut". It uses a working space which is defined by the user
...The internal working space of ACR and LR is Referred Input Medium Metric (RIMM), which has ProPhoto RGB primaries with linear gamma. Temporary excursions are made to other color spaces and image decompositions, as needed, to suit the image processing.
In ACR, there are additional options to set the desired output color space (which is used to drive the histogram and on-screen preview image). These color conversions are done at the end of the image processing pipeline, after the UI-driven controls (e.g., Exposure) are done. In LR, there are also such options (e.g., in Export and Web), though they do not drive the displayed histogram, nor the on-screen preview image. The color conversions are handled in exactly the same way...
-
I think they're quite appropriate. The question is, are they practical?
somehow all other ( != LR/ACR ) raw converters are using ICC profiles as a starting point (and whatever code they need to make it work).
-
I shot fine art for reproduction for 15+ years. I used a SINAR 4X5 (vacumn back) with AM ED Nikon APO lenses. Voltage stabilized tungsten lighting, cross polarized in a blacked out studio and I used the standard GMB color checker. I ran the Sitte Tischer line myself and could tweek a bit there if needed. Large drum scans were run in L.A. by the best and presses were profiled. We got really, really close, but certain colors were never really true.
Pastels - tuff
Water colors - pretty easy
Acylics - artist seldom cared
Oil - a bitch
Today I use a SONY A850 and CZ glass with the same lighting setup. We start much closer to the end result we're looking for and the time saved in the camera is now used in PS. Capture 1 does the best for most everything. I still run through ACR and a couple others when I get a new client but it's rare that we use anything but C1.
In the end it's still all subjective.
-
Re: Bill T's comment
I like to use the third grey patch down underneath the letters "ecker." Then go on and adjust exposure, brightness, contrast, and then go on to color tweaks. Incidentally the black textured case of the Passport is very nice for getting the shadow areas and overall contrast right.
I use Passport too not always happy with colours too saturated.
I was told to use the second grey from the bottom .. Why do you use third grey patch from the top???
This is one of the most interesting discussions for me in a long time as I only work with original art of all types for my wife and partner and many other artists, oil, pastels, material, rock dirt and whatever else artist find. They want prints to look the same ... not improved (unless it helps sales)
Would love a permanent category on this discussion forum for original art issues.
cheers elo
-
Elo Laugesen.
Of interest to me is what software do you use for printing original art reproductions?
Photoshop, lightroom, Photoshop Elements, Aperture, etc?
I have for years used Photoshop Elements as I did not need all the options and features. However the latest versions of Elements have been consumerised all the printing options eliminated, downgraded and I must now move on.....
-
Well, bought the Passport Colorchecker and just tried it out. Picked a painting with a lot of purples / magentas. Shot the chart, shot artwork under exact same conditions, built profile, applied profile, applied white balance.
On the training video the lady on the videos builds a profile and then applies it and says, "when you apply the profile you'll see THOSE COLORS ARE JUST POPPING RIGHT OUT!" Then she goes on to say, "You'll see it gives you a much richer, much more elegant look to your images." lol...what the hell is she talking about?!? The purpose of color management is to get accurate repro of colors in a scene, not to "make colors pop out" or appear "richer, more elegant." Good luck communicating color management using those terms.
So, I was a little harsh on my first review, so I've edited this. I've played with with the Colorchecker Passport. It's OK. Maybe a slight advantage, but I can't say it's any huge advantage over Adobe's presets for repro of artwork.
Are you using ACR set to ProPhotoRGB 16bits (and same for Photoshop working space)? If you happen to be using it in sRGB space, sRGB space might be clipping away the deep purples and magentas. Deep purples can be tricky, even AdobeRGB may clip them. I think you 221W display is 96% AdobeRGB those types of wide gamuts often both fall a little short of AdobeRGB and don't add all that much beyond the 98-100% AdobeRGB screens tend to add a good chunk of extras colors down in the purples and magentas and reds. If the painting has way intense magentas and purples they might be beyond even what your screen can show. Do you get any sort of image out of gamut warnings if you proof to your monitor space?
Sometimes CC makes a big difference but plenty of times it is more subtle. For the canned profiles from Adobe and Canon I find that Faithful is often the best (but CC is almost always at least a trace better and sometimes much better).
Sometimes monitor calibration probes are a touch off too. And then there is the issue of metamerism where a probe may measure same on two different things and they eye sees them differently. You'd need spectral distribution color management to avoid that and nobody inthe consumer world does that, at least not yet.
-
The painting I have chosen is watercolor landscape with lots of hues of magentas and blues. I shoot with two 3200K lamps on either side of the artwork in a blacked out studio, as I have for years. Shot the painting. Then shot the colorchecker. Built my profile per instructions. Did it again just to be sure. Then white balanced and saved as an Adobe Raw Preset. Applied to painting. Everything was too magenta. So I built another preset and white balanced using the white balance squares for color adjustment on the Colorchecker, essentially making the next preset I built "cooler." But, nay, now a little too cool. The blues are overtaking the magentas. So, I'm getting close, but already at this point the passport colorchecker offers no real savings in accuracy or time over just using the Adobe raw present for Camera Neutral or sometimes Camera Standard. I usually get them right after this many proofs anyway.
I'm probably going to return this. Honestly, if it was priced where it should be, I don't know, $20-$30, I might keep it for general use on portraits or something. Not sure if it would offer any big advantage there over Adobe's presets either. For painting repro, $100 for ballpark color accuracy is too much...I can do ballpark color with my eyes as fast as this.
Glad it works for you. If it doesn't help me match color in 3-4 tries or less, it's useless for me because I can usually get my colors right in 3-4 tries just using my eyes. I'll give it a try on a less challenging painting, but honestly, it's the challenging paintings I bought it for.
One thing is also that when you are white balancing it you are making it look as it would under D65 but in your studio you were looking at the painting under 3200K which is quite a long ways off from D65 and too far for vision to fully adapt to so I'm not sure a WB to D65 would be expected to match your view of the painting under 3200K.
-
Not really. See:http://www.color.org/ICC_white_paper_20_Digital_photography_color_management_basics.pdf
but in his case the goal IS to get accurate reproduction and in his sort of scenario you do not have any of the stuff that paper there goes on about, he does want direct colorimetric matching (when it comes to his having to make prints though, if he does, then it gets tricky since you will have to squeeze the gamut onto his printers and there are many methods and true testing is time consuming and tricky and not much work has compared different methods and various method can work better at various times and then his viewers need to view the print under expected conditions and not having just stared at a green light for an hour or anything weird hah, etc.
-
I do fine art reproduction for a living and have found that the Passport profiles are a little too vivid, plus the 24 samples are not very evenly distributed in the color spectrum, so there are some pretty big holes that the software has to guess at (I know MacBeth's motivation for choosing the colors they chose, but it doesn't help me much). I've had better luck with the Qp card, and their free Qpcalibration software. The 35 colors in the Qp chart are more helpful for art reproduction purposes, at least for me, and I find that I start with a closer approximation using that system than what I was using with a MacBeth chart and DNGProfiler (which I liked better than the Passport software). Photo printing is so much easier to do than art reproduction, just because the standards are so high when side by side comparisons are as easy as they are with art.
I will have a look at the Qp thingy. It certainly doesn't hurt to have more patches, even 35 is not so many really. And speaking of that yeah I wonder if 35 is enough to be worth the upgrade price. It maybe better to go at least CC SG.
-
2. Shoot RAW but do not use ACR for RAW conversion it does not have a wide enough gamut to accommodate all the colours you may need to preserve, it is also built for speed not accuracy detail smoothing and demosaic errors will appear. I use a heavily modified version of Rawtherapee that includes slow but accurate demosaicing my own input and output spaces tuned to fine art and is floating point for greater accuracy and detail.
How does ACR not have a wide enough gamut? You can set it to ProPhotoRGB 16bit for final result and I think they present using Melissa (ProPhotoRGB primaries with TRC of gamma 2.2) while using ProPhotoRGB primaries with TRC gamma 1 while calculating so various slider work better applied there than after conversion to gamma 2.2 or sRGB TRC or what not).
-
Roscolo:
What lighting are you evaluating the print and the original art work with?
-
This is where we were not long ago, so everybody cheer up.
Check the non-linear, marvelously crossed-over gray scale and the rich play of color casts across the matte surface. Back in the day there were a lot more just like this, and NO PHOTOSHOP and NO PROFILES. You kids today have it too easy.
Roscolo, let us know if you try that QP thing, or if you find some other repro-magic.
(http://lcweb2.loc.gov/service/pnp/cph/3j00000/3j00000/3j00100/3j00123r.jpg)
-
LOL, Bill. I sometimes get 4" x 5" chromes from clients that were shot by someone who got plenty of mileage from his color charts. The first color swatch on the left is labeled, Blue, BTW. Nice separation between the Magenta and Red swatches, too. (Note to self: don't store any color charts in a window sill or on the dash of my car)
-
Hello,
For a demo of color reproduction using a digital camera all the way to hard copy proofing, I started using Adobe's DNG Profile Editor to create a .dcp profile from a Colorchecker I shot. Next, I used Rags Gardner ACR calibration scripts (see here: http://www.rags-int-inc.com/ ) to really fine tune ACR settings like contrast, hue, brightness etc, get delta E's down. These values, together with the raw profile, I then applied to another picture I shot under the same lighting conditions. I then converted the raw file to ProPhoto RGB, and then absolute colorimetric to the printer profile. The printer profile had also been tuned with an iterative process using EFI Colorproof XF.
I have enclosed a jpg version of some real objects next to their reproduced counterparts. I'm not saying that every colour is perfect (let's say less than 1 delta E 2000), but this was the closest I could get using dcp profiles.
Mind you, it's not only reproducing colors found in fine art that is challenging. Think of photographing samples of wood panel floor (parquet), carpets, bricks, all for the purpose of offset-printing a catalogue with colors that need to be as close as possible to the original. Many commercial studio photographers I have met are struggling with this kind of work, and sadly more often than not, it's trial and error.
Regards,
Yann
-
Thanks for sharing that exercise and for pointing out those scripts and techniques.
I'm amazed a web image can make it to my Firefox screen with that much accuracy. For my particular evaluation regimen, if I slightly brighten the image in PS those charts are pretty much spot-on. There's a little blue contamination in the grey background here and there, maybe form a window or nearby object. But remarkable overall.
-
Roscolo:
What lighting are you evaluating the print and the original art work with?
Sorry I have neglected this thread. I evaluate print and original art under daylight (as in actual daylight), daylight balanced fluorescent, and 3200K lamps. I lean towards the 3200K lamps because most of the artwork and prints will be displayed under tungsten lighting conditions.
Used to use daylight tubes exclusively, but had enough scenarios where I relied on that lighting to make prints, then delivered prints and client looked at the prints in their home or studio under tungsten lighting where they will display them and they didn't always look as good.
-
Sorry I have neglected this thread. I evaluate print and original art under daylight (as in actual daylight), daylight balanced fluorescent, and 3200K lamps. I lean towards the 3200K lamps because most of the artwork and prints will be displayed under tungsten lighting conditions.
Used to use daylight tubes exclusively, but had enough scenarios where I relied on that lighting to make prints, then delivered prints and client looked at the prints in their home or studio under tungsten lighting where they will display them and they didn't always look as good.
If you're not editing your images so they appear pleasing under a particular light, then you're editing according to how the image looks on a neutral looking (6500K) calibrated display. You should never be editing your images for any particular light.
If prints match your calibrated display under one particular light (daylight fluorescent) and don't look correct under 2300K tungsten, something's wrong that can't be fixed by you, unless your client wants to not be so picky.
I've never had to edit my images for any particular light. My eyes quickly adapt to seeing the print under any given light except mercury vapor street lights which nothing looks good under.
The only color I can imagine not looking right under tungsten would have to be yellows, maybe purples, blues and cyans and where fresh, vibrant grass greens may look a bit cooked. But the warmth of the tungsten actually makes the print look better for these colors in my experience due to color constancy optical effects with WB color cast.
Below is an image I shot of a Skittles bag with my DSLR taken under all the lights I have and view my prints under. The Skittles bag was specifically chosen for its intense yellow to clearly show the spectral reflectance holes each artificial light brings out. As you can see the yellow looks a lot better under the GE Soft White tungsten (even better than actual daylight). It's a desirable color error. But this image is to really show how it's almost impossible to get a print to look the same under any given light unless you view them side by side under each light and that's not how we view prints.
The Smithsonian isn't concerned about these color errors viewing 100+ year old fine art painting under their MR16 tungstens, so why should your client?
-
As your illustration demonstrates, there is a huge difference between 6500K and 2800K household tungsten. We're talking blue vs. orange. Absolutely changes the appearance of artwork and prints and anything else that reflects light. Like I said, I try to look at prints vs. originals in both lighting conditions, but these days, for fine art repro, I tend to compare the prints I'm making against the originals under tungsten 3200K lighting as most of the work I am producing will be displayed in a gallery, museum or home environment under tungsten light. With fluorescents more common in homes now, that is changing for the home environment. My clients aren't picky so much as they want the print to match the artwork under the lighting in their environment. I do have many clients who want this level of accuracy, including a state museum. Hence my disappointment in the Colorchecker. I've always used just my own eyes, and my eyes are better than the Colorchecker. My expectations for the Colorchecker were perhaps a bit high. I have ordered the QP Card, though, thanks to the recommendation on this thread, and I'll report back how it does vs. the Colorchecker.
I'm an architectural photographer by trade. And I've operated a gallery for the last 20 years. I didn't really set out to be a fine art giclee' printer, but somehow the other two trades have somehow combined that I have gotten exponentially more and more clients for fine art repro, and those clients seem to have a higher and higher demand for accuracy. I stand corrected. You're right. They're damn picky! :)
-
I've always used just my own eyes, and my eyes are better than the Colorchecker.
I try to look at prints vs. originals in both lighting conditions, but these days, for fine art repro, I tend to compare the prints I'm making against the originals under tungsten 3200K lighting
Inkjet ink on paper and paint or any other media in a fine art piece can have completely different spectral reflectance characteristics especially under 3200K tungsten lighting that make it impossible for a color target to fix all possible color perception errors. Note I said perception which isn't how the color on those targets are measured. The eyes are very forgiving when it comes to looking at the overall appearance of any object or scene lit by any particular light.
Also to consider is how would one know if it's the light, the paint/ink or our eyes causing the errors? And what's considered an error if the entire piece looks pleasing but a spectrophotometer indicates huge Delta E number discrepancies. Do we go with what's pleasing? or accurate according to the numbers?
Note the lemon yellow turning cadmium yellow under the tungsten lights of the Skittles shots. Which yellow is the correct one? Who defines it? Our eyes? or some spectrophotometer? What if we don't like what the spectrophotometer says it should look like according to the numbers? The X-rite Color Checker patch pigments are manufactured to exhibit stable, spectrally flat colors under a D50 light source as measured by a spectrophotometer. Fine art paint isn't.
You'ld have to create color target that exhibits every possible spectral reflectance induced color error compounded by both tungsten light and any odd spectral properties inherent in the pigments, suspension medium and substrate of any given fine art piece.
IMO I think you're better off fixing each color error by eye as you encounter with each fine art piece viewed under tungsten because I doubt you're going to have trouble with all the colors in all the pieces. But if this QP card gets you closer, then it may be worth the effort.
Look forward to seeing what you find.
-
There's something else that came to mind that wasn't touched upon and that's the gamut of the inkjet ink compared to the fine art paint when viewed under tungsten.
The Skittles demo points to this issue even viewing on an sRGB display compared to a wide gamut display. I can tell you the way the Skittles yellow appears under the GE Soft Whites is a lot more intense than what I'm seeing viewing the image inline on my sRGB display in this thread. Because my eyes adapted to trying to match this intensity I thought I got pretty close, but after coming back, it isn't as intense as I'ld thought. I actually maxed out the yellow saturation in ACR on the Raw image. A dual illuminant camera profile didn't help either.
If you have paints in any of the fine art pieces that gain in similar intensity under tungsten, you may find your printer inks may be the limiting factor in which you'll have to know when to rule that out so you don't spin your wheels.
-
Thanks for that excellent demonstration photo. You should make a poster, would sell in droves to giclee makers.
Yes the spectral reflectance thing is one of the big bugaboos, not to mention the tendency towards metamerism which that and other effects bring to inkjet prints. And there's this newly recognized kid on the block called "subsurface scattering" which is a close cousin to spectral reflectance.
Hey Roscolo, I'm feeling kinda poorly about having pitched that Colorchecker to you. It works for my wife's very punchy pastel work (yes it's possible!) but have never tried it with subtle colors, or in situations where it's necessary to reproduce the subtle surface qualities and shading of the original media on the giclee media.
I just tried out the Datacolor "SpyderCHECKR" which is a sort of giant sized Passport. It's got the exact same patches in a different arrangement, plus an additional 24 patches in much more subtle shades. And yes, those deep-blue and crimson-red patches that the Devil put there are still present.
It's more trouble to use. You have crop it tightly in Lightroom, then send that image to the Datacolor software. But what it does is not create a DNG, but rather a lightroom preset with the controls in the HSL panel and elsewhere all tweaked around. You need to have the same Lightroom profile applied as when you made the preset. I have to say the match looked not too good last night under tungsten only, but this morning with sunlight coming through the windows it looks pretty decent, relevant to the tlooknbill was saying. Would NOT want to say this is "the solution" but it's an interesting approach. It also gives you a choice between what it calls "Saturation" and "Colorimetric" intents, Colorimetric seems to be more subtle.
I have also found that with the X-rite Passport, I can considerably get away from the slight oversaturation by playing with just the Hue and Luminance controls on just the red channel and magenta channels. I think the X-rite Passport might be giving us a Saturation intent, rather than a colorimetric one, but I see no way to specify anything like that in the software.
Also, I bought one of the QPCard201's from BH Photo and the darned thing only works through .jpg's! Good grief, what's that about! You apparently need the 203 model (or something) to work at the RAW, and nobody in the US seems to carry it. Sigh.
-
Also, I bought one of the QPCard201's from BH Photo and the darned thing only works through .jpg's! Good grief, what's that about! You apparently need the 203 model (or something) to work at the RAW, and nobody in the US seems to carry it. Sigh.
You can order it straight from their website and it's less expensive that way also. The 201 is JPEG only. I ordered the 202 card and one of their neutral gray targets, QP 101. The QP 202 is larger and supposedly a little more sensitive than the smaller cards. The QP 202 is Only $45. Much more reasonable than the $90 - $100 passport Colorchecker. Check out their plug-ins for editing camera profiles. Some of the links don't work unless you create an account, but creating an account just entails putting in your name, address and an email address. There is a menu on the right for changing the prices from Euros to Dollars. And if you do an order, the tax is removed at the last step of the ordering process.
I'll follow up here when I get the thing as I'm going to try it on some of the same paintings I used the Colorchecker on. Once you know the Colorchecker's default is grossly oversaturated and know to correct for it, the Colorchecker is a little more useful. The level of accuracy I'm having to produce, no matter what I use, I will usually be having to go in and adjust individual areas of paintings. And that's not really such a big deal. Again, I didn't really ask for all this business in this area, but I usually go in the direction the customers and the $$$ guide me to.
Most of these customers have had prints made at online giclee' printers and other places, and these artists aren't getting the level of accuracy they need. So I've gotten tremendous word-of-mouth business in this area. I'm always able to get pretty spot-on accuracy, but with all the extra business, I just need to speed up the time / labor it takes to go from first proof to perfection.
-
Here's a smartphone pano of a gallery space. 6300K on the right, 2600K to the left, all on the same wall. It's every giclee maker's worst nightmare and the bane of every artist and the norm for most galleries with a window at one end.
In taking pictures of installations in my customers' homes I often see interior walls in the 2600K to 2800K range. That's the same color temp as sodium lighting, but of course with a more complete but still highly biased spectrum. What can we do about these situations? The answer...get used to them.
-
One of my worst nightmares is hearing the ‘term’ giclee. ;-)
Our eyes and visual system would treat the scene you illustrate quite differently than the smartphone (or any camera system).
-
Here's a smartphone pano of a gallery space. 6300K on the right, 2600K to the left, all on the same wall.
OFF TOPIC
Bill,
If only the photo was a raw file, it looks like a perfect image to try out the localised colour temperature adjustment in LR4 Beta! I wonder how well it would work on the JPEG - when I have installed the beta (currently doing backups before a necessary OS upgrade) I may give it a try.
Regards
Nigel
-
One of my worst nightmares is hearing the ‘term’ giclee. ;-)
Our eyes and visual system would treat the scene you illustrate quite differently than the smartphone (or any camera system).
And that is generally the case at the gallery, the eye does not register either blue or yellow walls even when it can see both areas at the same time.
HOWEVER, I have to insist that the overall psyhological effect of say an image with either a markedly cool or warm ambiance is different over in the cool window light to the right, versus in the warm light to the left. It's pretty dramatic in some cases. And of course most of the print colors look different in the two extremes, although very few viewers pick that up directly.
-
It looks as if LED's are improving enough to please museum curators as shown here...
http://www.hogarthlighting.co.uk/acatalog/ironmongershall.html
I'm a bit skeptical LEDs could make paintings look that good since they have far worse spectral reflectance qualities than some daylight balanced fluorescents. Right now they're just too expensive for me to try them out.
-
The 3 or so different brands of "gallery quality" LED lights I have actually seen in use have had all the subtlety of disco lights. Certain narrow ranges of colors in inkjet prints pop out like dayglow paint, others simply fade to murky grunge.
Anybody using those Hogarth lights? The light looks decent so far as a web shot can represent. As far as I can tell about $600+ to light a sofa sized piece.
I'll be seeing stuff like that tomorrow at a framing show in Las Vegas. Will report if anything looks promising. A good LED light source would certainly be welcome, electricity costs can be pretty stiff for a properly lighted gallery, not to mention challenging the air conditioning in summer (although helping the heating in winter). But, cost would have to come down to no more than about $150 per large piece of artwork before any major adoptions at retail level.
-
i found that using adobe's dng profile maker is a great timesaver. i shoot the color target and the dng profilemaker application has a tab labeled 'calibration'. you put four cross hairs over your photo of the target and it generates a color profile recipe automatically. the next step is to export the profile. dng is smart enough to export it to the correct folder of adobe camera raw so that when you open a raw image in LR or ACR in PS, the profile is right there. i do find that the target leads to more saturated colors.
-
Update: Ordered the QP Card on Jan. 27th. As of Feb. 15th, I still have not received, and the QP company does not respond to emails. May have to take this to dispute with my credit card company. Even coming from Sweden, this is a long time. I will update if this company shows any interest in responding to my emails or letting me know the status of my order. At this time, I would NOT recommend ordering a QP Card from QPCard.com.
-
Update: Ordered the QP Card on Jan. 27th. As of Feb. 15th, I still have not received, and the QP company does not respond to emails. May have to take this to dispute with my credit card company. Even coming from Sweden, this is a long time. I will update if this company shows any interest in responding to my emails or letting me know the status of my order. At this time, I would NOT recommend ordering a QP Card from QPCard.com.
why not just call them - I guess all people in Sweden speak English...
-
Update:
Never received my order. They responded to my email and sent my order again, and included a free add'l. smaller QP color card to compensate for the hassle. So, that has now arrived. Looking forward to using it, and will update here. I'll probably start a new thread as this is QP, not Passport.
-
I do fine art reproduction for a living and have found that the Passport profiles are a little too vivid, plus the 24 samples are not very evenly distributed in the color spectrum, so there are some pretty big holes that the software has to guess at (I know MacBeth's motivation for choosing the colors they chose, but it doesn't help me much). I've had better luck with the Qp card, and their free Qpcalibration software. The 35 colors in the Qp chart are more helpful for art reproduction purposes, at least for me, and I find that I start with a closer approximation using that system than what I was using with a MacBeth chart and DNGProfiler (which I liked better than the Passport software). Photo printing is so much easier to do than art reproduction, just because the standards are so high when side by side comparisons are as easy as they are with art.
Received the QP Card, and they even updated the software for me (that's what I call product support :) ) because I had an issue with my camera profile not being "seen" by Photoshop (camera model name issue). So I have not used QP extensively, but on my first piece, a challenging watercolor, the QP Card camera profile was almost 100% perfect on the first proof! I should point out that the QP system has several settings, so you can build a profile using about 9 different presets, and it was one of these 9 that was dead on. Way better than the Passport! You'll likely have to order it from Sweden, but it is reasonably priced and worth the wait. Highly recommend the QP card!
HUGE thanks to Colorwave for the recommendation!
-
Happy to hear that I didn't lead you astray, Roscolo.
-
I'm curious why there hasn't been more discussion of the limited gamut of the paper/ink combinations as a 'culprit' in the lack of visual match. Someone, forget who, mentioned it but then it just went away.
-
But how does the QP Card 203 compare to the ColorChecker SG?
I've used Qualux Vopho, it is really good but the color is a bit too saturated, which you can easily re-adj. the setting under the software and recreate a new one after your test
Aaron
-
One problem I have with the Color Checher is its 4 x 6 grid format. In order to incorporate it into an image, I frequently have to pull back the camera. This can waste shed loads of pixels that would otherwise be capturing the artwork.
Ideally, I would like a 1 x 24 grid CC, available in large & small squares. In short, similar to the Kodak strips we all used in film days.
For me it is imperative that every image contains a chart since the files can often be used separately afterwards.
But, I would also like to be able to maximise resolution.
Great thread, by the way.
D.
-
Sent an email to QPcard to ask about their product referring to discussion on this site.
specifically I asked about capturing the right data for original artwork as compared to photos...
below is their second email... I will be buying one of their charts as I find passport colours too bright/gaudy in colours.. ( or is it me?)
Hi!
Lars already wrote a few things, but I though that some things were maybe still a bit unclear.
1) 203 is most probably better for you, since the 202 is slightly more sensitive to specular reflections. The 202 IS larger and easier to incorporate in larger works though, so as long as you have a reasonable control over lighting angles (and I guess you do, since you're into art-repro) 202 should not be a problem in your case.
2) The exposure has to meet those requirements:
-no colors overexposed in the raw file. When you open a file, no "blinkies" should be seen in the raw converter
-not severely underexposed. If you can add more than +2Ev exposure without getting blowing the "white" patch, you've probably lost some profiling accuracy.
There's a +/- 1Ev "window" of optimal exposure for the software to work with, and generally this is not very hard to satisfy. It WORKS with -3Ev exposure too, but the result is less accurate.
For your type of work (art-repro, of which I have done my fair bit...) often the exposure curve tends to mess things up a lot. The calibration gives you "accurate color" when everything is set to "zero", i.e no added contrast, no s-curve in the shadows and so on. When you add those things in, you oversaturate the midtones - where the exposure curve of a normal S-shaped type is at its' steepest.
steep curve - midtones - saturation increases
shallow curve - shadows and highlights - saturation decreases
When the RGB values has "one leg in the highlights and one leg in the midtones" you get a hue shift.
All of those things has to be adjusted for with human interaction, since there's no "perfect" exposure curve for any given type of art. But if you know the effect things have on your colors, they are quite easy to correct for. And when you have one good "oil, with semi-matte surface structure" setup, then this works for most pieces of similar qualities, regardless of artwork color constitution.
Feel free to as any more questions you might have, I will help you with setup of the calibration (via email of course...) :-) if you want some hints.
Greetings from Sweden
/Joakim Bengtsson
Senior developer, QPcard.com
-
I shot fine art for reproduction for 15+ years. I used a SINAR 4X5 (vacumn back) with AM ED Nikon APO lenses. Voltage stabilized tungsten lighting, cross polarized in a blacked out studio and I used the standard GMB color checker. I ran the Sitte Tischer line myself and could tweek a bit there if needed. Large drum scans were run in L.A. by the best and presses were profiled. We got really, really close, but certain colors were never really true.
Pastels - tuff
Water colors - pretty easy
Acylics - artist seldom cared
Oil - a bitch
Today I use a SONY A850 and CZ glass with the same lighting setup. We start much closer to the end result we're looking for and the time saved in the camera is now used in PS. Capture 1 does the best for most everything. I still run through ACR and a couple others when I get a new client but it's rare that we use anything but C1.
In the end it's still all subjective.
Amen to everything you wrote above.
I use a multi-shot Hasselblad. MS files can only be processed using Hasselblad's own software. I use this set to what they call 'Reproduction Mode'. Straight out of the box colour is extremely good.
White balance is achieved using a http://www.basiccolor.de/basiccolor-gray-card/ Fantastic piece of kit!
Just wondering if anyone has made test comparisons of Adobe's Profile Editor vs. X-Rite's, using Passport for both? I also shoot art using a Canon 1DsMklll, so I'd be interested to hear.
Lastly, the input of the designer and the printer (and his knowledge of his printing press) cannot be underestimated. Best printing I've ever had done of my work was in Brussels. Don't ask me why.
Thanks.
D.
Ps. Just ran a test to compare Adobe Profile creator vs. X-Rite as follows:
Evenly lit image of large X-Rite and basiccolor grey card.
Created profile in both software products, then opened them in ACR. I then made a white balance using the grey card, set the white patch at 245, the black patch at 53/54 and adjusted the brightness slider to make the grey card 143/143/143 (its correct A1998 value).
This image was shot using the 1DsMklll and the blue and yellow patches are always a problem IMHO. In Adobe profile, blue is 44, 60, 139. In X-Rite it's 32, 56, 139. Not bad all round.
In Adobe, yellow is 246, 216, 58. In X-Rite it's 250, 217, 41. Again not bad all round.
The green patch was 104, 156, 81 (Adobe, very good) and 110, 166, 83, X-Rite - less good.
Finally, viewing the chart in a D65 lightbox, the Adobe created profile looks less saturated overall and more accuate than the X-Rite one.
-
Hence my disappointment in the Colorchecker. I've always used just my own eyes, and my eyes are better than the Colorchecker. My expectations for the Colorchecker were perhaps a bit high. I have ordered the QP Card, though, thanks to the recommendation on this thread, and I'll report back how it does vs. the Colorchecker.
Have to disagree. Used with ACR (LR, percentage readouts and Melissa (FFS!) RGB are a complete joke for serious colour work in my opinion - Prosumer, NOT Pro) and it 0-255 Adobe 1998 readout, I think that a CC has to be more accurate than the naked eye.
That said, if your clients are happy....... :)
D.
-
IMO I think you're better off fixing each color error by eye as you encounter with each fine art piece viewed under tungsten because I doubt you're going to have trouble with all the colors in all the pieces. But if this QP card gets you closer, then it may be worth the effort.
In an ideal world, yes. But, most of us who do a lot of this find ourselves in a situation where art is put in front of the camera and removed immediately after the shot. The client wants as much art as possible documented in the day/half day they've booked you for.
So, the only practical solution is to rely on accuracy of the captured ones and zeros back at the office, and hope that the designer and printer follow suit up the line.
D.