UV Printer - Ink restriction for max Gammut

[virtu]

Hello,

It says in Profiling guide that i should set mit Ink restriction where the color doesnt bleed and no artifcting is visible.

So i printed the Ink Restriction Swatch....

I could restrict my colors at about 90 %...measuring density....but then my Cyan looks like a darkblue, my Magenta looks quite red and the Yellow looks like some 0/40/100/0
If i'd restrict around 60% the primary colors look like real C,M,Y,K

I dont wanna Profile for swop or euroscale but for max gammut. Will my Profile still look ok if i restrict that high like 90% ?

Thanks for help.

virtu

[walter_cfl]

use more patches for any charts you print and read, use Ultra Accuracy patches for ICC profiles.

[jonkovach]

I would think that if you are oversaturating your inks to begin with, that no matter how many patches or what accuracy level you use (I have found Standard accuracy to be more reliable than Ultra accuracy, anyway) - that you are going to have incorrect colors later in the profiling process.

Am I wrong?

[walter_cfl]

well, trust the new Onyx engine, give it a shot.

if still not comfortable, you could always try restrict your max ink base on LCH model instead of Density, look for the "hook" on C axis.

happy hunting!

[virtu]

Hello,

I created my own calibration swatch (20 swatches) and scaled it to 138 inch. Then i measured each swatch several times. I haven't done this before, i think it was worth.
Should have used 31 patches, will try that next time.

I didnt use the high Accuracy swatch but the ECI2001-cmyk target...it has about 1700 swatches i think.

My profil is quite good now, i match 50% spotcolors with an 4C printer.

Next time i have to do following better:
use more kalibration swatches, use high accuracy target.

Thanks for help.
virtu

[walter_cfl]

sounds like you're not using the ProductionHouse's profile generator, anyway, remember to save your profile as Version 4.

[virtu]

I heard Photoshop has troubles with softproofing version 4 profiles?

No i dont use the Onyx Engine for building the profile. When i started making my first profiles i noticed that my profiles were more accurate if i use EyeOne Match. I didnt know much about profiling , maybe i just set wrong settings in Onyx, Eyeone Match didnt allow too much "wrong" settings.

I havent tried the Onyx Engine since that. But my next Profile is 31 calibration swatches+ ultra accuracy chart+linearization.

Thanks,
virtu

[jonkovach]

I profile using Onyx, and I have found that I get better, more natural (is that the right word?) colors using the Standard accuracy as opposed to Extreme, or Ultra, accuracy. Has anyone else experienced this, or am I out in left field, spinning in circles?

[walter_cfl]

in virtu's case, the engine needs as much info as possible from the charts to map color on the gray axis correctly, remember, cyan looks like blue, magenta looks like red, yellow looks like orange.....

well, if you set your ink restriction based on the convention way, CMYK looks like CMYK, standard accuracy should work just fine.

just ensure your charts printed correctly, no defect, read correctly or your gray will run into trouble regardless standard or ultra-accuracy.

hope this explained a bit.

[virtu]

Hello,

It sounds logical that more swatches in ultra accuracy swatch produce more accurate results.

But what i dont understand is, why is it recommended to use only 19-31 swatches for calibration? If i understand the calibration process correctly, then i print lets say 31 patches and measure them, based on these measurements i create linearization.

In the wiki i read that more then 31 patches can cause a not "smooth" curve and spikes. But my machine prints exactly like that. If i measured lets say 100 PATCHES, then my densitiy curve was exact, shouldnt the linearization be based on exact values?

I hope my english was good enough so you understand what i mean

thx, virtu

[walter_cfl]

Agreed, 100 patches are good if it gives you "smooth curve", but it takes lots of time and material, not to mention not all substrate could give you the same result, substrate surface condition, dot gain, nozzle condition, how steady is your hand if you using eye-one, all this add up.....

virtu, you're a brave man

[virtu]

I hope my boss buys this 1 day:
http://www.barbierielectronic.com/p_spectroLFP.htm

Dot gain and surface condition are a factor too. I saw a demo of the spectro LFP live. I asked the operator if he could measure the same color 10 times, the value was indeed 10x the same. Well, he measured a HQ-print, some Epson or Canon printed Chart.

Dunno how accurate it was if his chart was some 300dpi print.

I dont think i will use 100 patches maybe i try 60 first, but i found out that measuring densities in Spot-mode is far more accurate then in Stripe-mode (dont get errors+neagative values).

I wont giveup that quickyl

virtu

[walter_cfl]

the Spectro Swing should be cheaper and provide your both reflective and transmissive reading, if, you do not print on thick substrate, i got lucky, using this Swing for whole full week, doing profiles for HP L65500 and Z6100, it works.

i suggest you verify your output profile by printing your inkjet proof agains a analog proof prints, this should give you an idea how things work.

[jens]

and if you are cutting to deep in to the ink limits , you are cutting in your color gamut. 4pass and 8 pass also make a great diffrence.