- This topic has 3 replies, 2 voices, and was last updated 2 years, 8 months ago by
.
-
Posts
-
Hi Florian, hi everyone.
Here the goal: I have to order a fabric providing CMYK values to the shop. The fabric colour should be as similar as possible to a the colour of a specific wall.
My idea is to use DisplayCAL + i1Display Pro to measure the RGB values of the wall, then converting RGB to CMYK for the shop.
Problem is that I really don’t know how to use SW and HW to do that (I know how to calibrate a monitor, not how to find RGB values of an opaque surface…)
I suppose I have to place the i1Display on a tripod at about 1 meter from the wall, and the wall has to be illuminated with a D65 light (the backlight of my monitor).
1. Then, how to obtain the RGB values from DisplayCAL?
2. Since I want to obtain CMYK values for a print on a fabric, is that correct illuminating the wall with D65, or should I use a D50 light? Or – since both wall and curtain will be illuminated by the same D65 light source in my room – should I measure the wall with D65 and then “conform” anyhow the obtained value?
(I’m clearly a newbie, so sorry for this level of questions.)Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Don’t spend time to alchemistry, believe me, it’s hard to measure print color correctly with colorimeter. Though, you may use the same colorimeter to find equality between two different screen colors under the same lights. To go in this way, start measurement process (white point manual correction stage) in DisplayCAL, holding i1d on the same laptop screen, then aim its lens to the wall. Note: the same colorimeter instance, the same computer screen to start measurement, the same room lighting, the same luminance force. And black cloth protecting colorimeter and light bulb both of side wall and furniture color.
But printing house can’t use measurement got from colorimeter. You may hire i1Pro spectrophotometer (test it with official i1Diagnostics & X-rite Device Services driver) and use “spotread -v -H” to read L*a*b color coordinates (for D50 WP, it’s familiar to press industry), use also “-S” if you want to see color spectrum. But this doesn’t resolve all your problems too. Color perception depends on ambient light, so if you want to see the same color under LED light, then printer ICC profile have to be computed for this lighting to get real equality. Be sure, your printing house can’t do it. Most of RIP software don’t allow to use user’s lighting spectrum for computing ICC profiles. My ColorGATE PS21 allows either to use some kind of standard aged bulbs, or to adopt image color for ambient light with 3×3 matrix that is rude enough way. You can measure and save ambient light spectrum with i1Pro. But… but… used i1Pro are not precise enough to get clearly the same color, i1Pro 2 are quite better, i1Pro 3 are simply used a little. Printing machines are not very accurate in color reproduction, paper production don’t provide stable tint.
So, direct comparison with some sample is the best way to find and print right color.
Dear Aleksey, thanks for your reply.
You are strongly right: comparing with some fabric samples would be the way to go. But unfortunately the printer is in Canada and I’m in Italy…
Just curious about theory and method: let say the printer use D50 illuminant, and let say I use Nix Pro 2 (native is D50 as well) to measure my wall; now let say I receive the fabric, I put it side by side the wall and I illuminate both with D65 LEDs… Theoretically I should see both (wall and fabric) with the same colour, right? (Theoretically…)
Or not?Nix Pro 2 seems to measure XYZ or Lab coordinates, not full spectrum. Perhaps it is ideal. Now avoid OBA-containing fabric, avoid inks thats spectral curves are away enough of standards for paints, use high-CRI LEDs or high-CRI CFL / tungsten / natural lighting.
I disasembled i1Display 3/Pro colorimeter, it has 3 grey-colored square photosensors (photodiods or like that) and 3 colored filters: blue, yellow and red. No green filter. I suppose, filters colors are close to L/M/S human eye cones spectra, that are not actually RGB, we get RGB after normalization. So, colorimeters are very sensitive to metameric effect and they’re tuned up to standard viewing conditions and standard RGB screen emition. But they also need correction for actual screen RGB spectrum and… consumer models are not precise devices, instances are different and tend to aging, so you need personal correction for your colorimeter and some display. DisplayCAL can do it for Spyders and i1Display, but data from spectrophotometer are needed.
That means, I doubt in Nix Pro 2 capability. It looks like a phone-photocamera with high-CRI LEDs lighting and special software. Hope, it is realy well calibrated for D50 and standard conditions measurement. Try to compare its results with i1Pro 2/3 spectrophotometer on “natural” paper prints. Standard targets like ColorChecker 24/Classic may not be very precise and wears out easily, fresh IT8 on photochemical paper with batch/personal reference data seems to be better, but has small patches. And print house measurent devices / profiles quality / print machine aging and working state / inks stability / supplied fabric batch difference ruin belief in equal color by measurement. (Though you’ll get photo very close to calibrated display, cause overall delta E 2.5 is well enough for print.)
The last problem is our mind attention to near white tints and large sheets comparison. I don’t know how much close the resulting colour should be for your task. Probably, all is OK.
Hm, can you send your Nix device to print house? Probably, it is the most solid solution. Another solution is color distribution proof: order a small piece of fabric with different samples / print tuning, print house should hold the same piece to avoid production error).
