HELP! How do I adjust for Delta E variances in Photoshop?

[jmyers732003]

I am trying to color match a printed sample. I have my readings from the Epson Proof which is profiled to the press. The digital image is in Lab color.
The differences are:
Delta L = 0.32
Delta a = 4.22
Delta b = -2.87
Delta E = 2.86

These readings are taken at a midtone color.

The customer is trying to get below a Delta of 1.5. How can I adjust the image in Photoshop to get the delta closer to a 1.5?
Any help would be greatly appreciated!

 

[thebestcpu]

Hi @jmyers732003
Some clarification and extra information would be helpful in giving good suggestions.

Note that ideally, if your work is fully color managed from you digital image, to your monitor, to your proof printer, or to the final printer, you should be able to create close to the exact color you desire on output from applying the desired color number in Photoshop. There are exceptions to this if you are printing to a printer whose paper and ink combination does not suport the desired color (e.g. it was out of gamut) and the rendering intent option shifts the color on printing (e.g. perceptual).

It would also be good to know the color mode and color spaces used for the entire workflow. You can used Lab color mode in Photoshop yet when printing almost all printers either need RGB or CMYK and an associated color profile with each. The chosen rendering intent make a difference as well.

Is the delta numbers comparing a reading off the customers sample compared to a local Epson print sample? What device is being used for the measurement?
Can you provided the actual measurement numbers for both as opposed to just the delta between the two. e.g. from just a delta number I don't know if the Lab numbers are higher of lower from the customer target print.

If we are not working with a complete color managed and calibrated workflow is one path and if we are just winging it to nudge color numbers close then that is a totally different approach (and harder because it could take some iterations)

Answering the above will make any suggestions I or other forum members make more on target for a better desired result.
Hope this will be a good start to resolution
John Wheeler

 

[jmyers732003]

John,
I have the Epson and Presses calibrated to CMYK G7 standards. My monitor has not been calibrated to the presses (yet).
The readings were taken from the customer supplied sample (target) and from our Epson proof. The device is an X-rite Exact spectrophotometer.
The art is a continuous tone image.

one sample reading:
target:
L=47.07
a=-26.22
b=14.85

proof:
L=52.89
a=-26.21
b=11.55

E=6.13
we need an E of 1.5 or less

The confusion I am having is that I was able to supply the Delta variances to the artist and they were able to adjust only with the Delta variances. Now, I have to go in and adjust myself but unsure how to adjust Lab vaule according to a Delta variance. I thought there was a way to do that in Photoshop...

I am used to adjusting the CMYK curves which I am most likely going to have to do because I am on a time schedule to get these ready to print.

Hope this helps John!

 

[thebestcpu]

Hi @jmyers732003
I was on vacation and just got back.

My interpretation of your situation is that for some reason which is probably not important, the color calibrated workflow did not provide a match to the final press print and you want to "nudge" the images color to be closer to the final target.

I will note that working with a single color on the target and the proof make get you closer yet how much the color shifts at higher and lower luminosity may not track as there are many ways to adjust colors. So this may not be a single shot process.

To "nudge" the colors vs having a calibrated and measurable color numbers all the way through the workflow, I would adjust each of the "L" "a" and "b" numbers by the delta amount needed to go from the proof to the target.

I have shown in the table below the numbers you provided as well as the deltas for each component in 8 bit and also in 16 bit. (16 bit is just multiplying L by 256 and a,b numbers by 128.

	L	a	b
Target 8bit / 16bit	47.07 / 12,050	-26.22 / -3,356	14.85 / 1901
Proof 8bit / 16bit	52.89 / 13540	-26.21 / -3355	11.55 / 1478
Delta from Proof to Target 8bit / 16bit	-5.82 / -1490	-0.01 / -1	3.3 / 423

So here are the steps I would take.

Convert you image to 16 bit mode
Convert to Lab mode
Above the image Layer place a curves adjustment Layer
In the image place a Color Sampler on the part of the image you want to measure when "nudging" and change the info numbers to diplay in 16 bit mode not 8 bit mode being the default
Open the info panel so you can track the color change.
Since you indicated the color measured was mid tone I would make the changes in the cures adjustment Layer at the middle of the curve line.
With the L channel selected in the curves adjustment layer dropdown, move it down while watching the info panel until the new L number drops by 1490
I would not bother with the a channel since the delta is minuscule
With the "b" channel selected in the curves adjustment layer dropdown, move the curve from the middle position upwards again watching the info panel until the "b" number increase by 423.

After this you can change to the desired color mode and then to 8 bit if needed.
It is important that the process of outputting/printing is exactly the same as when you did the proof because it is unlcear where in the workflow that the error in the color managed workflow exists.

I tried this out using two layers with each with the Target and Proof Lab numbers you mentioned and easily adjusted the single Proof color over to the Target color using this technique.

If you need more details or step by step with pictures I could do this yet I was under the impression you may have moved on already and was not sure it was worth a lot more to post details unless you needed them

John Wheeler