For an accurate color chart, what is the best method for "capturing" the colors?

[HarveyDunn]

For an accurate color chart, what is the best method for "capturing" the colors?

I have a client who wants "the most accurate possible" printed color chart for a line of 89 artists paint colors. They are made with 'traditional" pigments, which means that none of them are blindingly bright. I'm fairly certain they will all be within the GRACOL profile's gamut.

I have real samples of each (paint thickly applied to canvas, low gloss/matte surface, nice and flat application - no ridges or brushstrokes.

I'm wondering what the best technique would be to capture those colors and transmit them to InDesign. The obvious choices seem to be:
*photograph them
*scan them (desktop flatbed?)
*measure them with a spectrophotometer, perhaps multiple times and average the results. Input the LAB values into InDesign and let the Adobe color engine convert it to CMYK

Any thoughts about which of these is best? My gut says it is the last one, but the only spectrophotometer I have access to is an i1 (or is it Eye One). Is that good enough for the job?

 

[gordo]

I have a client who wants "the most accurate possible" printed color chart for a line of 89 artists paint colors. They are made with 'traditional" pigments, which means that none of them are blindingly bright. I'm fairly certain they will all be within the GRACOL profile's gamut.

I have real samples of each (paint thickly applied to canvas, low gloss/matte surface, nice and flat application - no ridges or brushstrokes.

I'm wondering what the best technique would be to capture those colors and transmit them to InDesign. The obvious choices seem to be:
*photograph them
*scan them (desktop flatbed?)
*measure them with a spectrophotometer, perhaps multiple times and average the results. Input the LAB values into InDesign and let the Adobe color engine convert it to CMYK

Any thoughts about which of these is best? My gut says it is the last one, but the only spectrophotometer I have access to is an i1 (or is it Eye One). Is that good enough for the job?

First off '"the most accurate possible" printed color chart' isn't going to happen. Even if the colors are in gamut for CMYK process the spectral properties of the pigments are different as is the whiteness, brightness and OB content of the printed chart vs the artist's substrate. That means that the hues will shift differently under different lighting conditions. And with artist's supply retailers you have zero control of lighting.
So, you could do what house paint manufacturers do and actually print swatches of paint to make charts (this is also what Pantone does for their inks). Or you can do some of the things that artist paint suppliers do:
Print CMYK simulations to make a printed swatch chart but also instruct the retailer to paint actual swatch samples on canvas for in-store display. Or, the actual paint swatch samples can be created separately and included with the printed charts. I.e. One actual paint swatch chart for retailer display included with 1,000 cmyk printed charts.
Use actual paint to daub on the label of the paint tube.

If you measure the color of a paint sample to get its Lab color make sure that the paint surface is flat and use a spectro to measure the color rather than do a scan or take a photo. The paint manufacturer may already have Lab values for the paints for their own QC requirements. Some paints will be semi-transparent or transparent so the surface that you apply them to should have a similar OB content to what the paint is typically applied to (e.g. a gesso ground). You'll need a profile of the print condition of the setup that you will be using to print the final charts in order to get reasonable CMYK recipes. Use maximum GCG screen tint builds in order to minimize the effect of the spectral differences and to help stabilize the color on press. Use an FM screen if possible to eliminate moiré, rosettes, and graininess.

 

[HarveyDunn]

Thanks. These are for a mailing to end users, not in-store use. The manufacturer does not have the LAB values so I'll have to get them myself. Do you think an X-Rite i1 is up to the job? What is "OB content"? Whare are "GCG screen tint builds"? Lets assume this is going to a G7 qualified printer and I've got the GRACoL profile set up in color management in InDesign. Can I just make process swatches using the LAB values and rely on the Adobe Color Engine to do the rest?

 

[gordo]

Thanks. These are for a mailing to end users, not in-store use. The manufacturer does not have the LAB values so I'll have to get them myself. Do you think an X-Rite i1 is up to the job? What is "OB content"? Whare are "GCG screen tint builds"? Lets assume this is going to a G7 qualified printer and I've got the GRACoL profile set up in color management in InDesign. Can I just make process swatches using the LAB values and rely on the Adobe Color Engine to do the rest?

1- an X-Rite i1 is up to the job

2- OB content or OBA content is the optical brightening agents used in most paper to increase the apparent brightness and whiteness of papers. They do this by absorbing energy in the ultra violet and emitting (fluoresce) the energy in the blue area of the visible spectrum. Because, to the eye, blue/white looks "whiter" than yellow/white OBAs are not really whiteners, but bluing agents. Their effect on the appearance of the paper (and hence the hue of the transparent ink) is affected by the amount of UV in the light source. More info is here: The Print Guide: The issues of Optical Brightening Agents in paper and ink

3- "GCG screen tint builds" is a typo. ;-) it should have been GCR screen tint builds. Whenever 3 chromatic (CMY) colors are combined - one of them actually just greys (neutrally darkens) the color. So, typically in a separation some portion of that chromatic color mix can be replaced with a percentage of black (achromatic) ink. One reason why you have CMYK separations rather than just CMY separations. The amount of CMY color that is replaced by black is selectable. Most separations are done with Medium GCR (grey component replacement) applied. Using Maximum GCR increases the amount of chromatic color that is replaced by Black. The less there is of chromatic color then the less that color can shift on press or under different lighting.

4- If you're using a generic GRACoL profile then it may not represent what your particular printer will actually produce when printing the swatches even if they are a so-called G7 qualified printer. Especially if the color swatches are just flat patches of color. There is enough variability within established tolerances for printing to make some color patches shift in hue while being within specification.
It all comes down to what you really mean by "the most accurate possible printed color chart."
If it really is that critical then the best way is to do an actual proper press test that includes "ring-around" screen tint builds - i.e. a patch of what you think is the correct screen tint build surrounded by patches that vary one of the color screen tints.
So you really need to set the expectations of your client by establishing what your client means by "the most accurate possible printed color chart" because that will have a direct connection with production costs. Greater accuracy = greater costs.

 

[spoocobra]

Back me up Gordo (hopefully) if you wanted to truly get the best accurate representation couldn't you get the LAB readings from scanning them, talk to your printer and have them custom mix inks for each color (like spot) and then run a print test with so called "ring-arounds". Although this surely will be an extremely expensive and time consuming endeavor. Best bet i.m.o. is to find a G7 Master Printer, get your LAB, RGB callouts and make a big swatch board, export with out ANY icc profiles and give to your printer and work very closely with them and their proofing system they have. I Don't understand how they don't the manufacture doesn't have LAB or some kind of callouts for these inks?

 

[gordo]

Back me up Gordo (hopefully) if you wanted to truly get the best accurate representation couldn't you get the LAB readings from scanning them, talk to your printer and have them custom mix inks for each color (like spot) and then run a print test with so called "ring-arounds". Although this surely will be an extremely expensive and time consuming endeavor. Best bet i.m.o. is to find a G7 Master Printer, get your LAB, RGB callouts and make a big swatch board, export with out ANY icc profiles and give to your printer and work very closely with them and their proofing system they have. I Don't understand how they don't the manufacture doesn't have LAB or some kind of callouts for these inks?

I seriously doubt that a scan of a patch of paint will give you a Lab value that would be of any practical production value. Spectrophotometers are designed to do that and their ability to do that is reflected in their cost. If the selected printer has a printed CMYK atlas (The Print Guide: The Color Atlas - helping designers to specify color) then that could be used to provide 4/C recipes if there are samples of the paint.
The printshop does not have to be a G7 Master Printer to provide those services. They do need to have a repeatable standardized process though. AFAIK paint manufacturers use recipes (like Pantone does with ink) to define their colors. Although CIE Lab values, being numbers, seem absolute and unambiguous - in fact they may not be (The Print Guide: Top reasons why color instruments don't agree).
So maybe I got your back up rather than backing you up :-(

 

[spoocobra]

I seriously doubt that a scan of a patch of paint will give you a Lab value that would be of any practical production value. Spectrophotometers are designed to do that and their ability to do that is reflected in their cost. If the selected printer has a printed CMYK atlas (The Print Guide: The Color Atlas - helping designers to specify color) then that could be used to provide 4/C recipes if there are samples of the paint.
The printshop does not have to be a G7 Master Printer to provide those services. They do need to have a repeatable standardized process though. AFAIK paint manufacturers use recipes (like Pantone does with ink) to define their colors. Although CIE Lab values, being numbers, seem absolute and unambiguous - in fact they may not be (The Print Guide: Top reasons why color instruments don't agree).
So maybe I got your back up rather than backing you up :-(

Well when I said scan I meant with a spectrophotometer of course lol, but wouldn't/couldn't you in essence use a pantone or swatch book like you posted to, to "visually" match that paint samples as well and just make a big board and go from there as well?

 

[gordo]

Well when I said scan I meant with a spectrophotometer of course lol, but wouldn't/couldn't you in essence use a pantone or swatch book like you posted to, to "visually" match that paint samples as well and just make a big board and go from there as well?

Oooops! My apologies - I misunderstood what you meant by "scan" - methinks you "measure" with a spectro. ;-)
Ideally the printshop would have their own color atlas. I wouldn't use Pantone's spot color swatches nor trust their CMYK screen tint builds to represent what any given printer would deliver.

 

[spoocobra]

Oooops! My apologies - I misunderstood what you meant by "scan" - methinks you "measure" with a spectro. ;-)
Ideally the printshop would have their own color atlas. I wouldn't use Pantone's spot color swatches nor trust their CMYK screen tint builds to represent what any given printer would deliver.

Scan/Measure, Potato/Potatoe lol. Do you feel getting the paint manufacture directly involved with the printer as well will help in any way?