425978
The exchange platform for all who are interested in newspaper printing
New ICC profile for newspaper production
New ICC profile for newspaper production
Submitted by Anand Srinivasan on Thu, 2015-04-16 15:04
Forums:
The ISO 12647-3 standard for newspaper production is updated in the end of 2013. A new ICC profile is being tested.
Are there any upcoming training events or tutorials announced in the area of color managemnet, color profiles and Color Quality Club?
For the benefit of the publishers, who wants to participate in the Global print quality competition, INCQC 2016-18, we have planned many public trainings in different part of the world.
1. INCQC Tutorial, 15 June, Frankfurt, Germany. Language: German
2. INCQC Tutorial, 7 July, Dubai, UAE. Language: English
3. INCQC Tutorial, 13 July, Chennai, India. Language: English
4. INCQC Tutorial, 15 July, New Delhi, India. Language: English
5. Color Quality and INCQC Tutorial, 29-30 July, Kuala Lumpur, Malaysia. Language: English
6. Color Quality and INCQC Tutorial, 5-6 August, Rio De Janeiro, Brazil. Language: Portuguese and English
For registrations, please contact anand.srinivasan@wan-ifra.org
WAN-IFRA, jointly with a group of Swedish printers, has developed a new ICC profile for newspaper production (beta version). The recent revision of the ISO 12647-3 standard has reduced the recommendation for TIC from 240 to 220%. It necessitates a change in the ICC profile.
This profile was tested today in Malayala Manorama, one of the largest newspaper houses in India.
Test Methodolody: Same RGB images were converted to CMYK using the ISOnewspaper26V4 profile and the new WAN-IFRA beta profile. The images were placed top and bottom to be in the same ink zone and printed. The paper and ink used were ISO 12647-3 compliant.
Observation:
1. Both the profiles are perfectly compatible with each other. For an untrained eye, both pictures would look exactly the same.
2. On close observation, it is noted that the saturation of the picture is marginally high in the new profile. It is more noticeable in the red regions. Skin tones gets little bit more red.
3. The increased saturation also leads to loss of contrast in those regions where saturation increases.
4. The shadows are more open due to the reduced TAC. Shadow details are hence improved.
5. Slight reduction in print through was observed. Wherever there are dark shadows, it appeared lesser in the back side compared to the ISO profile.
All the above observations were very subtle and it can easily go un-noticed.
The main objective was to create a new profile that is compatible with the previous ISO profile so that it can be implemented in production without major adjustments. Considering that, the profile is good to use.
Thanks to Malayala Manorama for offering to do the test run.
We did calibration on our newsapaper machines and we set profiles in GMG color server as given below;
1. We are converting files RGB into CMYK
2. After this we are directly dropping these files into ISO coated V2 Fogra 39L
3. Then we are converting from Fogra 39L to IFRA26
Is this method is ok, or what will happen if we directly drop files into IFRA26.
I observed flesh tones are wash out and highlight areas are becoming dull. Looking for experts advise
Dear Ponnapati,
At WAN-IFRA, we do not recommend this approach, where you convert an RGB image to a large CMYK color space (ISO Coated) and then re-convert it to smaller newspaper color space. The ISOnewspaper profile is an excellent, well-tested and used world-wide profile. Why not convert the images directly to the newspaper profile? (This profile will soon be replaced by another profile that is created as per the recommendations of ISO 12647-3: 2013. The new profile will be available from end of this month). The profile is created with maximum GCR and do not need any additional processing for ink savings as well.
When you convert an image from RGB (Larger color space) to newspaper CMYK (Smaller color space), there is a shift in colors as the smaller CMYK color gamut cannot accommodate all the colors from the larger RGB gamut. It is inevitable.
The solution lies in how you prepare your RGB images. When you do color correction, either manually or in automatic software, You prepare a "Heavier" RGB image. Then, when you convert the image to ISOnewspaper, it lightens it and you should get the colors to your liking. This fine tuning of your color correction process to suit the destination CMYK profile is part of any color management process.
If you are doing manual correction, it is simply done by using a CMYK preview during color correction (Short cut: Ctrl + Y). Place the right profiles in the color settings. Then, when you apply the CMYK preview, you see exactly how the image will appear after the RGB to CMYK conversion. However, the image will still in RGB. If you do the entire color correction with the CMYK preview, you can color correct to your liking and then on conversion, there is no visual change to the colors.
Using sRGB as your RGB color space will also help.
Dear Mr. Ponnapati,
first and frontmost mistake you did is that you have converted your (raw, original) RGB file, completely irrelevant at this point have you color corrected it completely to satisfy your needs and expectations or not while still in RGB color space, to a much larger CMYK color space than Newspaper26v5 (or previous v4 version) actually is. For start, your CMYK color profile in it's name has a word "Coated". That should tell you everything right from the start.
Now, that larger CMYK profile is being mainly used for "Magazine" or Heatset prints on a coated paper that can take and reproduce on itself much more colors, saturated, vivid, denser colors than our Newspaper is able to at all. It's gamut is much larger (wider) than Newspaper26 profile (v5 or v4). So, due to the re-mapping process while conversion occurs in your next step from one larger CMYK color space to a much, much narrower (smaller gamut) CMYK color space results in, of course, duller or "washed out" images with less contrast and much poorer colors (less saturated, less contrast, low "information" to be able to show in print) than what you get after conversion from RGB to your CMYK color space in your first step.
Correct way should be, if you work in Coldset newsprint industry, color manipulation (editing, retouching) of images while in RGB and then converting your images to Newspaper26 profile. Also, bear in mind very important thing which is correct Color Settings for Newspaper production, if working in Adobe Photoshop (look it up on web) with activated command "Color Proof" (Cmd+Y on Mac or Ctrl+Y on PC). That activated command "shows" you your RGB file how does it "looks" in CMYK that is set up in Color Settings for Newspaper Coldset print. Gives you a "preview" how does an image you're editing "looks" like in targeting Newspaper "world" even that you're still in RGB mode. Very last step should be, if working with final CMYK files, converting to Newspaper26 profile.
The other way is to leave an image in RGB which was corrected while "showing" you how will it look in CMYK using Newspaper profile but working with final, color corrected images in RGB requires different approach to the complete workflow in printing house but that's for another topic.
I hope now you have a clearer view what is the right approach in preparing images for Newspaper print.
Best regards,
Krešimir Bikić
Hi,
I try to use WAN-IFRAnewspaper26v5 to convert our RGB images to CMYK for our newsprint. I use Photoshop with Relative Col. (it's correct?) and black point compensation. I see the darkest area have 180% of total ink coverage, seems too low because in the ISO_12647-3-2013 i see the number must be 220%, 240% maximum. I ask it's normal when convert a black box RGB to CMYK with WAN-IFRAnewspaper26v5.icc profile (rel col and BPC), the TIC are 180%, or anything is wrong.
Best Regards!
Hi Dani, For RGB to CMYK conversion, the recommendation is to use Perceptual rendering intent without black point compensation. You would get a TIC of 216 with the new WAN-IFRA icc profile.
Relative colorimetry with Black point compensation would provide you minimal difference in colors, when converting from RGB to CMYK but it will also reduce the contrast of images. Please use Perceptual intent without BPC.
In a separate email, I will send you a "Color settings File", which you can implement in photoshop
Thank you Anand!! But I think somthing wrong. Please see the this random (and pretty!) image. At left the RGB original, front the CMYK file from your settings. The CMYK file is brighter (burned), loss detailed at skin, loss color at hair, etc. TIC 208% . At right in CMYK, v5 ICC, Relative Col. and BPC (more looks like the original RGB file, as you say). Tic 180%
If I use your setting, all my images burn when convert from RGB to CMYK. I think this is not correct. I do more test with others images and have the same result, images burned like our lovely Monica!
Hello Dani, I have been asked to comment on this topic and I have looked at your test picture. I have to congratulate, you have put our attention to two important topics, se below.
Method for CMYK separation, there are no right or wrong here, just differnt technics with some drawbacks. I usaly say that everybody who´s not experts on digital image handling ought to use Perceptual, and the drawbacks are that the image gets lighter and a bit more yellowish. But a big advantage is that the image gains more contrast and this is something that images in coldset often needs. Relative Col with BPC looks more like the original, and this is good if your original i perfect. But in most cases this is not the case and then the perceptual rendering intent is generally more forgiving.
First important topic on your image and the occurred problems. This is due to the fact that the RGB image don´t have a full histogram in the blackpoint. Any CMYK ICC-profile will always assume that we start from a full histogram with R=0, G=0, B=0 (no light, realy max black) and goes to R=255, G=255, B=255 (white with no tones). If you start at R=9, G=9, B=9 (like your image uploded) the ICC-profile don´t map this to max black in the CMYK-profile, it maps this a bit lighter black, as it ought to else your images would have heavy black areas with no tones (clipping in black areas). This effect becomes bigger if BPC is used with Perceptual rendering, se below.
Second important topic on your image, ICC-theory states as far as I know that BPC only ought to be active when using Relativ rendering but when testing this in Photoshop 5 and CC on ISOnewspaper26v4.icc and WAN-IFRAnewspaper26v5.icc BPC is active in Perceptuell as well. This is why you get the washed out result on your image when using Perceptuell rendering. If this is due to Adobe or Profile Maker and i1Profiler or if I just have missed something I dont know. But to summerise:
Perceptual (without BPC) = good, forgiving separations with more contrast.
Perceptual with BPC = not good, dull separations that loses contrast in dark areas.
Relative with BPC = very good if you have perfect images.
Please note that this behavior is present in ISOnewspaper_26v4.icc and just a bit more distinct in WAN-IFRAnewspaper26v5.icc
Please use the test image provided and experiment with different rendering intents, with and without BPC.
I made a PDF shoving separation methods and values in your picture, sorry but Photoshop is in swedish but I think you can understand anyway ;-)
Good luck with the tests.
Hi Magnus and thank you to explain and put more light!
In short, you said: 'Relative with BPC = very good if you have perfect images.' but when convert RGB to CMYK with Rel. and BPC, TIC are only 180%, and I can see in my test lost detail and color saturation in dark zones. Sure it's correct? the CMYK image must have 220% TIC (Ifra ISO v5 ICC), seems 180% is too low ink.
Backgound:
I need to use rel. col. with BPC because three important things. First, alot advertisement becomes in RGB and it's no good idea to over contrast when convert to CMYK. As you said, if the image are perfect in ad (or near), the best conversion is rel. col. and BPC.
Second, we work under a automatic image processor for our photo news (Q-Enhancer and Automata). Theoretically, the reult are a perfect retouched photo (you know perfect perfect...) but in RGB, next convert to CMYK using a ICC profile. Here cannot use v5 Perceptual without BPC because the image over contrast again, when image has been automatically retoched by QE.
And third, we create images for our work, our adv, etc. I know if now put the v5 with Perceptual and no BPC, our operators scary when convert to CMYK. The v5 over-contrast cannot adjust (more or less contrast) and some images goes well but in some others no.
OK! I see, each image needs own settings to convert to CMYK. But the problem it's our production itself. We need an unique ICC profile with the same settings for unify all prepress, and works to print, of course. The Holy Grail for our workflow may be Wan-Ifra v5 ICC profile with Rel. Col. and BPC but... TIC!! :P Total Ink Coverage with these settings are 180%, lower than 220%. May be 180% is good too, low ink like ink saving prepress servers but not sure, I see lost detail and color saturation... by the way, how I can have the result of a conversion to CMYK like rel. col. with BPC but the TIC goes up to 220%?? We don't have any bench to modify profiles, only Photoshop to create a custom CMYK one.
Im not sure to do, but for now:
V5 ICC - Perceptual - No BPC: Not for our workflow bcause wash every image
V5 ICC - Rel. Col. - BPC: Low TIC (180%)
Create a profile with Photoshop? I try to make one with settings (image). Ink colors are from Wan-Ifra info in LAB coor. and dot gain curve too from Ifra. It's a valid ICC profile? In an image, I see changes between these profile and Ifra v5 but when converting from RGB to CMYK TIC are 220% with Perceptual no BPC and 215% with col. rel and BPC (good)
Now I don't have more ideas but the safe method is to use a custom ICC profile to convert from RGB to CMYK.
Ifra ICC washes or low ink my images... see the comparison image. The only good conversion for me is with a custom profile, but may be only visually and not real print. Yes, test test and test again :)
Best Regards!!
Hi Dani,
I understand your concerns. You are trying to adapt a new ICC profile for your current way of working with RGB images and it is not working.
I used to do color correction for about 4 years, when I worked for a newspaper. We were using the older ISOnewspaper26V4.icc profile at that time. When working with photoshop, our team use the option CTRL + Y to simulate the CMYK conversion without actually converting the image. By this way, you will see the image as how it will appear when you convert it to CMYK. Manual color correction is done on this preview. We correct the images to our liking, convert it to CMYK and finally satisfied (We don't see any shift in colors as we work on the preview, where the colors are already shifted).
As Magnus mentioned, the choice of rendering intent is yours. Our team always prefered perceptual intent because the profile lightens the image. Most images that we receive needs to be lightened for newspaper production. Hence, using perceptual intent means lesser color correction steps. The corrected RGB image that we have is darker and when the ICC profile and perceptual intent is used, it is lightened to our liking.
Why the perceptual intent lightens the image - that is one of your question. The profile is not supposed to do color correction. You can try perceptual intent with other profiles like Euro standard or SWOP and you can see that perceptual intent does not lighten the image at all. It happens only for the newspaper profile. The newspaper color gamut is very small. Even when you use sRGB profile as your RGB work space, there are many out of gamut colors. When you use perceptual intent, it tries to map these out of gamut colors into the newspaper gamut and in the process, shifts all the colors. This results in the change in colors and the image becomes lighter . It also means that Perceptual intent is the best method to handle out of gamut colors. Colorimetric intent simply clips the colors
Now coming to your problem - You have two challenges
1. Your current color correction settings in Q-enhancer is suitable for the earlier CMYK profile + rendering Intent that you have been using and is not suitable for WAN-IFRA profile + perceptual intent. Solution is to adapt the Q-enhancer setting to suit the new profile and rendering intent. I simply cannot recommend relative colorimetry + BPC as photoshop is reducing the TIC to 180, which is too low. Adapting the color correction software settings to suit a new profile is a standard task that one has to do when changing the profile.
2. Advertisements - This is the biggest challenge. Ideal scenario is if the advertisers send you the materials color separated already with the WAN-IFRA profile. It does not work always. When you convert the ads to CMYK in your end, I agree that you have to ensure minimum difference in the colors. In this case, relative colorimetry with BPC will ensure less differences (Still, out of gamut colors will be clipped). For this scenario, I can recommend the older ISOnewspaper26V4 profile with BPC. It will give you a TIC of 215.
My recommendation for editorial pictures, where you have complete control, is to use the WAN-IFRA newspaper profile + Perceptual intent + Modified Q-enhancer settings. This profile with perceptual intent is used in hundreds of production houses worldwide and it works great.
I don't recommend creating and using a custom profile with Photoshop.
I really like this discussion, both Dani and Anand are doing the same mistake as I have ;-) Assuming that TIC 220 always is in the blackest area (R=0, G=0, B=0). But this is not the case, the profile is now "smarter" and will use the available CMYK colors where they are most needed, in saturated dark areas. Look at blue and black areas in the provided images. ISOnewspaper26v4.icc was poor at handling blue areas with relative colorimetric. Notice that blue areas turn red and how the pure channels get dirty. Look att the blue dress and how it looses contrast and gets flat... WAN-IFRAnewspaper26v5.icc is nearly perfect with relative colorimetric. Notice that blue areas get TIC of 218 and the black only gets TIC of 180.
Don´t worry about TIC 180 in the black areas, my qualified guess is that this won´t be visable in print.
Conclusion, always use the new profile. It is better at almost everything.Hi Magnus! you're right!!!
We uses v5 profile with Rel. Col. and BPC in our production and it's ok, but need to resolve other issues like printing process... a world apart!!
TIC 180: Black = C36 M21 Y25 K100 with the smart profile feature to keep color saturartion in dark areas... Anand (Ifra) agree that??
I attach a zip with some of our production images in CMYK v5, auto image retouch without hand manipulation. The images are OK to print? some advice?
Thanks a lot!!!