Yet another Ink Limit-Linearization question

[[email protected]]

EFI Colorproof handles this just fine as do some other rips. A rip like Colorburst Xproof now lacks what I consider an adequate process and I have a couple of interesting questions as well.

I am aware of the recommended process, 1. per channel ink limit -peak chroma or min L* change, 2. linearize channels, 3. set total ink limit.

A. Doesn't it make more sense to linearize again after setting total ink limit?

B. Would it be worthwhile to evaluate neutral grey of CMY at per channel limits -together with the total ink limit- and balance (tweak) the individual channel ink limits (a bit) to achieve a more neutral grey at the expense of a bit of saturation in a more saturated channel? Let's say the channel ink limit for magenta is saturated to the point where cyan and yellow can not combine with it for a neutral grey- rather than styep on the CMY channels equally with the total ink limit- how about taking down the Magenta channel a bit and then doing a total ink limit. You would lose a bit of saturation in the Magenta but you would gain a broader neutral grey scale (cyan and yellow will not be clipped as much by the total ink limit and will balance better against the magenta.)

C. Does it make sense to give the black channel ink limit a bit more preference if you are going to use higher ucr/gcr?

D. Recommendations for black settings for proofing profiles for an inkjet printer?

 

[rich apollo]

A) The RIP should be doing something like that

B) In my opinion, no. Inkjets inks are not the same as offset inks. You're looking for the best color gamut you can get out of the proofer. Gray balance will be handled during the conversion from the input color space to the proofer color space.

C) Not sure what you're asking here. You want all the channels to be effectively limited/linearized. You may find aggressive GCR to be counter-productive (unattractive) on an inkjet. The black ink doesn't serve the same structural purpose in inkjet printing. You may your inkjet black ink to be less neutral than your offset ink, and if your proofer doesn't have variable droplet capability, GCR will lead to "peppering".

D) I tend to run MUCH less aggressive GCR for inkjets - higher black start - less aggressive black generation, but that's my own preference.

 

[[email protected]]

re-lin and ink limits

re-lin and ink limits

Thanks for the reply Rich:

RE: A. I agree, the rip should relinearize after TIL. I don't think the one I mentioned is doing that - I will check and relin as necessary.

RE: B. I understand your response and I think that it is the widely accepted practice. I posed the question to see if anyone (besides me) thought there was a possibility of improving upon the balance between CMY values (not like a press- but improve.)

Say that after measuring individual ink channels, Magenta saturation at it's proposed channel limit is relatively strong compared to Cyan. We move on to total ink limit where Magenta and Cyan are scaled back in equal proportion. Magenta remains strong in comparison to Cyan. The imbalance may result in the device profile having to work hard to produce neutral tones and balance.

In contrast, at the point of individual channel limits, I could scale back Magenta slightly. Yes, that would reduce the maximum saturation of Magenta- However- when I do total ink limit- there is now more room in the TIL for Cyan and Yellow. Saturation of these inks will go up - thereby increasing the gamut. As a bonus the inks will have better balance and should make building an accurate profile more certain.

Thoughts?

 

[TerryWyse]

I am aware of the recommended process, 1. per channel ink limit -peak chroma or min L* change, 2. linearize channels, 3. set total ink limit.

A. Doesn't it make more sense to linearize again after setting total ink limit?

I assume you're talking about ColorBurst's in-RIP total ink limit and not ink limiting in the ICC profile...in any case, it won't make any difference.....linearization is pure CMYK ramps only whereas total ink is combined CMYK. Besides, CB's total ink limit is set AFTER linearization anyway. How CB sets their total ink limit is somewhat of a mystery since it's not based on a printed/measured target like it is with EFI's Color Manager......all I know is that it works well.

B. Would it be worthwhile to evaluate neutral grey of CMY at per channel limits -together with the total ink limit- and balance (tweak) the individual channel ink limits (a bit) to achieve a more neutral grey at the expense of a bit of saturation in a more saturated channel? Let's say the channel ink limit for magenta is saturated to the point where cyan and yellow can not combine with it for a neutral grey- rather than styep on the CMY channels equally with the total ink limit- how about taking down the Magenta channel a bit and then doing a total ink limit. You would lose a bit of saturation in the Magenta but you would gain a broader neutral grey scale (cyan and yellow will not be clipped as much by the total ink limit and will balance better against the magenta.)

In my opinion, no. I find that sometimes I'm in the minority on this opinion...but I don't think gray balancing an inkjet via either ink limiting or linearization is a great idea. You want maximum chroma (or nearly so) for the per-channel limits and a linearization that achieves optimum tonal distribution in 3D Lab space. In my opinion, giving a profiling application good ink limits and good tonal distribution of the primaries and secondaries will allow it to do it's job.....basically feed it good predictable data and it will do a good job of achieving gray balance and "mapping out" the color space of the device.

I actually went down the road once of trying to gray balance an inkjet and it was fruitless...from a 3D gamut standpoint, you'll end up with NON-linear data which is not a good thing.

C. Does it make sense to give the black channel ink limit a bit more preference if you are going to use higher ucr/gcr?

As long as you limit the black ink back from where the minL* plateaus, you should be good to go. I usually limit black about 95-98% of it's total L* range.

D. Recommendations for black settings for proofing profiles for an inkjet printer?

WHICH inkjet printer?
Rule of thumb....if it's an Epson Ultrachrome printer, start the black at 0% and go all the way to 100%....use 100% GCR and the heaviest K curve you can get away with. (This usually takes some analysis in ColorThink Pro so you don't bottom out one of the channels, usually yellow).

Other printers that use light K ink ("gray"), I would follow the same strategy.

If you're dealing with a printer with no light K inks such as the Epson 10600 and some of the older HP printers, then you have to be real careful with the K settings. Generally you will have to start the K no lower than about 30% and maybe as high as 60%. On these printers, you want just enough K to get a smooth result and good Dmax...but you have to avoid K "peppering". These printers are difficult to profile....usually takes several profiles to find the right combo of GCR, K start and K curve. Had one last week (an HP5500) that crashed i1Profiler.

Hope all this helps,
Terry

 

[[email protected]]

re CB ink limit and linearization

re CB ink limit and linearization

Thanks Terry:

Right now I am working with a 9600 it has light black, lc, lm. I built an awful profile following the CB process- looking at the neutral graph in Colorthink the Magenta and Yellow lines suddenly drop down to about 10% at about 80% black. The profile produces images with a moderate magenta color cast. I am pretty sure my poor results have to do with the channel and total ink limits.

I think CB is a great product but the new lin/ink limit process (introduced 1 to 2 years ago) doesn't seem to work well for me. The fact that it works well for you is hopeful news.

When pritning to a new paper-CB suggests choosing a similar paper from their environment list and re-linearizing- if the relin falls within their suggested values - they suggest using their environment's inklimits (channel and total) and profile (that is not reassuring to me.)

If the relin doesn't fall withtin the suggested values you can try moving channel ink limits (although not intuitive) to move it\them in to the suggested value range or you can start from scratch. When starting from scratch they suggest you print the linearization target- measure it and store it as base data and then print the profile target. If the target is too wet, then back off on the total ink limit.

Unless loading the measured linearization data accomplishes both channel ink limit and linearization, they are missing a step. Next you havee to set the TIL- you do this by printing the profile target and seeing if it is too wet. The process doesn't seem right to me.

any suggestions?

Thanks.

Kirk

 

[TerryWyse]

any suggestions?

I can give you a few quick suggestions....but if you want to some real detail, we'll have to do it "on the clock"

Any-who....

With your per-channel ink limits reset to 100% and with everything dispabled such as ICC, total ink and linearization, print a linearization target. If you have the PC version, you'll have to print the lin target from within the linearization dialogue.

After measuring the lin target in ColorPort/SpectralVision, open the measurment window and flip the little down arrow so you can measure the chroma values directly (Lch).

Find the peak chroma of each CMY step and note what dot% it occurs and what the chroma value is...should be in the range of around 70 for cyan, 80-84 for magenta, 105-110 for yellow. For most of the inks you'll find that the chroma will peak and then fall off as you approach 100% (yellow is usually the exception). Multiply the peak chroma by 95% if you want a more conservative ink limit, use 98% if you want to be more aggressive. So if the cyan peak was 70, 95% of that would be 66.5...now go back to the measurement window and find the step closest to 66.5 (less dot% than the peak chroma obviously)...note the dot% that gives you the closest value....do that for each CMY ink. For K, just see if you can find where the minL* value starts to plateau and use that.

Use the values you found and set those as your per-channel ink limits......in my experience you should print the lin target AGAIN and see if your target chroma values are now at 100% (my guess is they won't be)....keep pulling back the per-channel limit value until you get the value you arrived at in the first step. Once you do, use that as your final linearization and save the data and import that into the linearization dialogue and reset/set the baselin values....before you do that, you need to also set the lin method. I would avoid the chroma method....user either the Lab or Lightness method...I prefer Lightness.

At this point, I'll typically print the linearization target with linearization ENABLED and take a look at it....you should see nice smooth tonal transitions that look visually even. If all looks good, move on.

Set the total ink limit by simply having CB do it for you...leave the black level at 100% though....use the NEW total ink limit method, not the "legacy" method. You can check your total ink limit by printing one of their total ink limit targets but I usually skip it....if you've done the steps up to this point correctly, you should have minimal pooling/puddling.

Print your profiling testchart, measure it and make your profile using the suggestions I mentioned before (K running from 0-100%, 100% GCR and with a "Heavy" K curve. Total ink limit is somewhat a guess at this point but 300% is a good place to start...I usually end up around 240-260% with CB and most Epson printers on a decent proofing media.

At this point, you'd need ColorThink Pro to do a through analysis. If you have CTPro, then open your first profile and look at the neutral rendering curves. In rough terms, you want the cyan/magenta to roughly follow the K curve most of the way, at least through the midtones....the yellow curve will be lower but you don't want it hovering along just above the "floor" of the curve graph...get the yellow above about 20-30% at the midtone by going back and using a less aggressive K curve (don't touch the 0-100% limits, they're fine). Once you find the K curve that suits you, look at the minL* of the curves in CTPro.....with the rendering set to absolute, you want to keep the chroma value of the neutral curve <2.0...if it's higher than that, I would reduce the total ink limit when making the profile...do these reductions in about 20% increments until you find the setting that gives you a low chroma value AND a minL* you can live with....if you reduce the total ink limit too much, you'll see the minL* start to rise.

THAT'S the basics...and it's not exactly the techniques I use today....I have a whole new method that involves delving into the light/dark ink curves but it's not for the squeamish....I've developed a spreadsheet that does the hard math for me....but the techniques I outlined above SHOULD net you good results.

Terry

 

[[email protected]]

Solid Information

Solid Information

Thanks Terry,

I had the right idea but your specifics give me a clear path on getting there. I appreciate them and more so your willingness to contribute. Thanks.

Kirk