A few days ago we received a call from a customer in the field of design, who sent open Adobe InDesign data in ISOCoatedV2 300% with contained RGB images to the production company for a complex CD production on the advice of the producing company (“The printing company still has a prepress stage, which can then prepare your data optimally…”). The result of the finished printed CD booklets and inlays did not correspond at all to the calibrated monitor image of our customer, the client was also unhappy and requested the print data about the production company from the print shop responsible for the print to troubleshoot. Data in the “US Web Coated” color space with 350% ink coverage came back from the printer. For troubleshooting, the customer then had a proof of his data created by us, but had chosen the settings “Convert to target profile (retain values)” as usual when writing the proof PDF; we thus received completely CMYK data, of which we produced a proof according to ISOCoatedV2 300%, which completely met our customer’s expectations. So it seems that the designer created the data correctly and printed the print shop incorrectly.
On closer inspection, our error analysis revealed two serious weaknesses:
In this case, a complaint of the designer to the printing company will of course be difficult, as on the one hand, non-profiled RGB data were sent to the production company, and on the other hand, no print PDF generated by the data creator in the correct output color space ISOCoatedV2 300% was supplied.
If this had been done, one could at least have argued that the expected color of the production print would have been comprehensively known. Thus, one can only refer to the fact that the printer would have had to ask the designer for RGB data without an embedded color profile, and should not have assigned the data somehow to a profile “blindly”. The fact that the print shop with its crude US Web Coated workflow certainly did not create a correct print file, but a wrong one for the output, can indeed be stated, but the print shop can always talk its way out to “systems with in-house standard”.
If we receive a PDF file that contains RGB images, the next step is to check if the file is a valid PDF/X-3 or PDF/X-4. If this is the case, we check whether all input RGB profiles are correctly marked with color space (sRGB / AdobeRGB / ECI-RGB-V2 etc.) and rendering intent, then we check whether the correct output color space was used as output intent and whether also contained CMYK data have the correct input profiles. If yes, then we proof the file with the settings: “Consider all input and output color spaces”.
In this case, the file is reproduced 100% exactly as our customer created and defined the color profiles. If he has made a mistake and e.g. marked an image with a wrong RGB profile, this will also be “incorrectly proofed” exactly as correctly.
If RGB data should not contain a profile, e.g. if they are created in Device RGB, we generate a “data incorrect” e-mail in which we explain our procedure as follows:
“Dear customer, the data check has shown that RGB elements are contained in your data. RGB elements can only be safely interpreted in the proof if they are marked with a color profile and a rendering intent. This is the case, for example, with correct PDF/X-3 and PDF/X-4 data. The correct output intent must also be specified.
At least one of these criteria does not seem to be the case for your file. The safest way would be to convert the contained RGB data to CMYK. This has the advantage that you have control over the conversion and can view the CMYK result again in Acrobat before uploading the file again for proofing. We can then reliably use your CMYK values for the proof. To do this, call up the current order in your customer account, delete the incorrect data and upload the corrected data.
If, for example, the RGB element should only be a small image that is not relevant for the overall impression of the proof, or if you do not have another file available for the proof, then of course we can also use your RGB data for the proof. If available, we use your RGB source profiles and rendering intents, otherwise we use the sRGB standard and the rendering intent “relatively colorimetric with depth compensation”, which in most cases will lead to correct proof results. If you would like us to proof the supplied RGB data in this way, please let us know. Please do not hesitate to contact us if you have any questions. Best regards, your proofing team”.
In our case, the CD production case would also not have occurred in the proof, as we reject RGB data not provided with an ICC profile with the error message mentioned above, and do not convert them, as we cannot predict precisely how our customer would have liked the data to be converted.
We are aware that our approach is not 100% the ultimate best approach in all cases, but to the best of our knowledge and belief it is best in line with market practice and the expectations of our customers.
However, we are also happy to accept your individual requirements and circumstances. Give us a call or send us an email and describe your processing requirements.
By the way: We are happy to put our knowledge and data competence at your service: If you also have a problem, a question about print data, data preparation, or – as in the above example – a misprint has already occurred and you need external expertise and assistance for the complaint: Give us a call. We will be happy to advise you and help you where we can help. We will charge you for our advice and analysis at an hourly rate of EUR 90,- plus VAT, and you will be billed for 15 minutes each. An initial consultation and assessment is of course free of charge.
Due to our involvement with freeColour e.V., at the last meeting in Switzerland the desire for a cross-media tool for designers was expressed with which one can create intersections of colourspaces from the freieFarbe CIELAB HLC Colour Atlas XL.
With Gamutmap, Proof GmbH has now created such a tool, which is available to all designers free of charge. With Gamutmap nearly 100 individual colour spaces can be indicated from 34.250 colours of the entire CIELAB colour space, or intersections from many combined colour spaces can be indicated.
An example: As a designer you are looking for colours for a new corporate design, which are available in sRGB for the internet, in ISOCoatedV2 for printing image brochures and in PSOUncoatedV3 for printing stationery. For video productions, the Rec.709 colour space is also to be taken into account.
In Gamutmap you can now easily select the colour spaces sRGB, ISOCoatedV2, PSOUncoatedV3 and Rec.709 and then click on “show”. After a few seconds you will only see the colours that are available in all selected colour spaces. If you move the mouse over a colour field, you will directly see the absolute colorimetric values of the colour in all selected colour spaces and you can copy them directly to your clipboard.
Since the hex value of the sRGB colour space was also still interesting, this colour space was additionally marked for display. The HLC and Lab values of all colours can be read directly in the colour table. All other colour values can be copied to the clipboard simply by moving the mouse to the desired colour field. For the colour field shown in the example above, it looks like this:
HLC: H005 | L055 | C035
Lab: 55 | 34,867 | 3,05
sRGB: 188 | 106 | 128
sRGB (HEX): #BC6A80
Rec. ITU-R BT.709-5: 188 | 87 | 115
ISO Coated V2 (ECI): 14 | 64 | 27 | 11
PSO Uncoated V3 (Fogra52): 10 | 70 | 34 | 8
We are sure that gamutmap will be a great help to many designers in creating cross-media corporate designs and are very happy that we were able to start and push the project with the members of freieFarbe e.V. For us, gamutmap is “work in progress”, which means: In the coming weeks we will add further functionalities and features to gamutmap. For example, a German version is in progress, and the download of spectral D50 CxF data of the selected colours should be possible in the future directly while hovering over the respective colour field, if the field is in the gamut of the freefarbe CIELAB HLC Colour Atlas XL. Further function extensions are already on our wish list… 🙂
We welcome suggestions, criticism, wishes and any support for the expansion and addition of Gamutmap.
The question often arises whether color profiles should be embedded in the PDF files for proofing.
To answer the question, you have to get some answers: The proof should simulate the subsequent offset printing. For offset printing, with few exceptions, the imagesetters have been configured so that a 70% black in the file is displayed as 70% black on the printing plate, no matter what profile was specified in the file. It didn’t matter whether it was coated paper or uncoated paper: 70% in the file corresponded to 70% on the plate, the choice of the paper printed on resulted in the colour representation.
The proof has also adapted to this: Most proofing service providers ignore embedded profiles, as long as the data is in CMYK and do the same as their print colleagues. Even with grayscale, the profiles are usually ignored and the grayscale is simply assigned to the CMYK black channel. Thus all CMYK and grayscale data are simply interpreted as if they had been created in the output color space. If “ISOCoated V2” is proofed, all images are treated as such, and if “PSOUncoated” is proofed, then the CMYK images are created in this color space.
This is excellent for the majority of files to be proofed. Only RGB colors contained in the data are problematic.
Since the RGB color space is considerably larger than most CMYK color spaces, it must be clear from which color space to convert to CMYK according to which criteria. Most proofing service providers specify a color space from which they convert by default if no RGB color space is defined. This can lead to difficulties: For example, many proof studios choose AdobeRGB as color space because it is large and optimized for offset printing; however, most images from digital cameras come from sRGB and these color spaces differ considerably. Therefore, it is important that the RGB color space and the rendering intend is embedded for a proof, otherwise the proofing software normally selects a color space for conversion to the CMYK color space to be proofed; and this color space is possibly not the one in which the data has be created.
Colour is colour, you’d think. That’s right. But have you ever tried to explain the colour of your new car or your new red wallet to a friend on the phone? You will notice that human colour recognition and the reproduction of the same in another medium is very difficult.
The same applies to computers – better: monitors, and printers – i.e.: laser printers, inkjet printers or newspaper printing or offset brochure printing.
Why is the red on a monitor different from exactly the same red printed on paper? It’s simple: put the paper in front of the monitor. The two shades of red are exactly the same. Like this. And now you’re completely darkening the room. What do you see? The red on the monitor is still red. And exactly the same red on paper? This is black now. Why is that? Very simple:
A monitor adds light, i.e. spectral components, to the existing ambient light. If you see red on a monitor, it is because the monitor actively emits red light.
And now the paper: When do you see red on paper? Exactly when white light falls on the paper, for example through a window or a lamp. And when do you see the colour red on paper?
When white light falls on the paper and the paper extracts the non-red spectral components from the white light and reflects the red light. That’s when you see the colour red.
One colour, two completely different ways of production. And this is exactly where the colour calibration and the proof start. The strategy? Fairs. And this under fixed conditions and not with the human eye, but with “incorruptible” technology.
Put simply, a monitor calibration device can measure your monitor and see exactly “how much” colour your monitor can display, and “how wrong” your monitor can display colour. And if your computer knows that, it can correct the monitor.
Another measuring device can emit neutral white light onto a paper and measure the reflected colour. Depending on the printing process and paper, the ink looks completely different, but the meter again sees “how much” ink the print can represent and “how wrong” the print represents ink. And if your computer knows this, it can correct it. And:
If the computer knows the colour representation of the monitor and printer, it can correct and adjust the representation so that both correspond to the same colour. Of course, this only works if the colour and brightness of the light that illuminates the paper is also known and standardized.
And how does the proof work? Very simple:
If a computer also knows that the final printed product is to be printed in offset on an image printing paper, and it knows the colour representation of this printing process, then it can simulate this on a monitor and on an inkjet printer.
On the monitor, this colour-accurate representation is a so-called “soft proof”, the colour-accurate preview of the subsequent print on the inkjet printer is called “Proof” or “Contract Proof”.
This inkjet printing must be very precise and meet the highest demands in gamut and colour simulation. And since the image processing technology, colour matching calculation and measuring technology behind it is not very cheap, proofs are still mostly “expensive” inkjet prints. Due to new printing systems and inexpensive and better measuring technology, however, prices have also fallen significantly here in recent years.
Especially in larger companies today the layout in RGB is the rule rather than the exception. The advantages are obvious:
In practice, however, there are two potential problems in particular.
Problem 1: CMYK conversion in the last step.
The catalogue is designed in InDesign, all data is perfectly matched, the last step before printing and proofing is the export to a printable PDF in CMYK. Usually this is done via a preset in InDesign, which defines the exact specifications for the color space conversion. In practice, however, this color space transfer can hardly be monitored. The problem: Even if you check the color values in Acrobat in the exported PDF file, for example, Acrobat does not really display the colors it contains. Acrobat brav would show you CMYK values even if the RGB images are still wrongly contained. However, other CMYK values can occur during printing when the data is processed again. Lately it looked like this:
If you own Adobe Photoshop, you can do these conversions directly there. In Photoshop CC all well-known color books are stored with values.
Let’s assume we are looking for the Pantone equivalent and the matching CMYK color of HKS 43 K.
1: Open the color palette in Adobe Photoshop and select HKS K as the book and then the color HKS 43 K. All well-known colour books are directly stored in Photoshop.
Handling wrong profiles with CMYK data / “Profile Mismatch
If we have only received CMYK data from you, we will ignore all input and output profiles and only use the CMYK values that we bring to the ordered output colour space.
You send a file with the profile ISOCoated and a colour area in CMYK 100/70/0/0 and order a proof according to ISOCoatedV2.
We ignore the ISOCoated profile and proof the pure colour value 100/70/0/0 according to ISOCoatedV2.
Why do we do this?
In our proofs, we try to reproduce the “lived reality” of the print as well as possible. In many conversations with printers we have seen that in almost 100% of the cases they do not convert profiles from CMYK to CMYK, but instead put a colour value of 100/70/0/0 on the plate without taking CMYK profiles into account, insert paper and print in conformity with the standards. So we also map this way, although it would actually be “more correct” to perform a colour space transfer from ISOCoated 100/70/0/0 to ISOCoatedV2. However, this results in a different colour value, for example 100/63/1/6 for relatively colorimetric conversion with depth compensation or 100/63/3/15 perceptively with depth compensation!
One of our customers did not proof 30 slightly different, dark blue colour areas in ISOCoatedV2 on our premises, but on the premises of a colleague, under each of which the CMYK value was in black lettering, in order to sample the colour of a powder-coated surface. The customer defined a very well fitting CMYK colour value on the basis of the proofed colour areas, inserted it into his brochures and started the print jobs. Result: The dark blue was a distinctly different blue than on the reference proof, customer and agency were very dissatisfied and went on troubleshooting. Now the case came to us.
We received a file for proofing according to ISOCoatedV2 and compared it with our colleague’s proof. The colours with the same black CMYK values printed underneath were clearly different, but both proofs were provided with media wedges and measured correctly. After some troubleshooting, we came up with the idea of requesting the original proof from our colleague, which also existed. In this one there was a Fogra27Coated profile, thus an implementation of the old ISOCoated. A proof according to ISOCoatedV2 had been ordered at that time. Had it happened? The colleague had taken the input profiles into account, which resulted in a significant change in the CMYK values of the colour patches, as mentioned above, due to a colour space transfer from CMYK to CMYK. The black printed CMYK values under the colour patches had of course not changed. The patterned CMYK value therefore did not correspond to the proofed value at all. Our customer fell from all clouds: “How, our CMYK values were not proofed”. This would not have happened with us, because we would ignore the embedded profile with CMYK data. In this case this would also have been our customer’s expectations.
After almost two hours, we had determined the “error” (or perhaps rather: the “difference”), created a proof for our customer that was “in line with expectations”, which he could use to determine the appropriate CMYK value in ISOCoatedV2, and solved the problem. (more…)
This is the first incomplete list. You know of other good reasons? We look forward to any comments and would be happy to add further points.