Photoshop Custom Ink Values don’t match readings in document

[Lorna Wiles]

Custom Lab ink values are entered into Photoshop’s custom cmyk set up, these are then saved along with the other relevant settings as a custom profile. On using this profile either as the working space or by converting a document to it and then adding 100% of either C, M, Y or K the resulting Lab readings in the PS colour picker don’t match the original values entered in the CMYK setup, eg for Cyan:
Cyan custom ink setup values: L = 60, a = -30, B = -40
100% Cyan colour picker readings in the document: L = 64, a = -30, B = -45

Why is this?
Should I enter my actual ink values in the custom CMYK setup or in fact adjust these values before entering so that I get the correct readings for the 100% patches in the profiled document?

I'm sure there is an obvious explanation. If anyone could enlighten me it would be much appreciated!

Thanks

Jon

 

[Stephen Marsh]

The obvious explanation is not to expect accurate results from the legacy Custom CMYK engine. :]

Seriously though, the Custom CMYK engine can be useful, however it is not a “true” ICC profile tool. These days I prefer to avoid it where possible, although it is nice to know that it is there, even if the base technology predates Photoshop 5 (not CS5).


Stephen Marsh

 

[Lorna Wiles]

Thanks for the reply.

I somehow suspected that it was going to be the clunky cmyk engine!

However, in that the engines inaccuracies are consistent, ie we get the same result each time, is it better to put the true ink values into the custom cmyk setup or edit them before entering so that we end up with the true values in the document? (This would be the same for the secondary colours as well)

I also totally get the point that the colour engine is outdated and there are far better ways available for creating profiles. However I’m looking to try a simplified routine here as the screen printing press we are profiling has such a large fingerprint when compared to more conventional presses.

Thanks

 

[Stephen Marsh]

Lorna, I would not use the word “profiling” in the same post as “Photoshop Legacy Custom CMYK Engine”. This is not profiling (characterisation)!

How are you getting your measurements? How many patches are you reading? What are you measuring with? How repeatable is the process, does colour printing vary widely from job to job, material, screen or chemistry batches etc?

Use a remote profiling service to measure and create a proper profile from a supplied chart.

Get a proper spectro and profiling software.

Even get a ColorMunki – they are “inexpensive” and can create a better profile than kludging one via Photoshop’s Custom CMYK.


Stephen Marsh

Thanks for the reply.

I somehow suspected that it was going to be the clunky cmyk engine!

However, in that the engines inaccuracies are consistent, ie we get the same result each time, is it better to put the true ink values into the custom cmyk setup or edit them before entering so that we end up with the true values in the document? (This would be the same for the secondary colours as well)

I also totally get the point that the colour engine is outdated and there are far better ways available for creating profiles. However I’m looking to try a simplified routine here as the screen printing press we are profiling has such a large fingerprint when compared to more conventional presses.

Thanks

 

[Lorna Wiles]

Thanks for the reply.

The processes we are using can vary so much that it’s extremely hard to nail it down to a consistent standard.

The chemistry:
Although we mix / formulate our own cmyk inks from base pigments. Along with film production this is probably the most consistent part of the operation.

The screens;
This is the most variable part of the operation due to a multitude of reasons too many to list. It is however controlled as is practically possible.

The Material:
We always print on the same cloth. Although the white point of the cloth varies from batch to batch as dose the construction / absorbency, it falls within industry tolerances. These tolerances are far, far wider than any spec for paper would ever be.

Press
A press operator can expect to adjust a press for at least the first twenty minuets before it will stabilise as well as make adjustments during a print run. Lots of variables here that an experienced press operator can use to good advantage.

Targets and Instruments
We have had profiles made for us out of house (remote service) using the IT8.7-3 and i1 ECI targets. In house we use an SP68 (designed for reading textiles) print large targets (up to 18” x 36”) and measure the patches by hand - average of 10 readings per patch. Even how many threads there actually are per inch in the cloth where the patch is printed / measured is a variable and effects the light reflected from the sub straight under the cloth (I use a neutral grey)

With all the variables the only reason I think we can ever acheive any acceptable cmyk print is because our reproducible gamut, by flexo standards, is very limited, ie the fewer colours there are to reproduce the easier it is to reproduce them.

After 12 years of trying to nail a standard I’m beginning to think it is somewhat futile, the only reason for continuing the quest is of course professional pride - we can always do better. However when jobs fail even after remaking screens etc. I suspect it has more to do with image type, black generation and the black generation settings locked in the custom profile we are using.
My reasoning for using Photoshop was two fold, one to try and absorb all the variables. There are only 8 patches + a white to read. (A ‘phone enquiry to X-rite a number of years ago lead me to believe that fewer patches may, for us, produce better results). The second reason was for altering / experimenting with the black generation based on the image being prepared for print, ie is it dark or light, high contrast / low contrast etc.

So, to conclude do you still think I’m heading down the wrong road with Photoshop? Would a ColourMunki be better?
Other than give up, (which for me isn’t an option!) any other advice?

Thanks for you time.

Jon

 

[Stephen Marsh]

Thank you for the explanation Jon, it helps to know where you have been and where you are going.

It sounds like there are variables in production that account for colour variation. It can be pointless trying to profile a moving target, however some target may be useful compared to no target.

I thought that you were trying to avoid the cost of colour profiling hardware/software and that is why you were looking at Photoshop Custom CMYK as an alternative.

Measurement data can be made into many different profile types using different GCR etc, so once you have “accurate” measurement data (you may need to average multiple charts) it is easy enough to build a range of profiles to suit different work.

You might be able to get usable results from Custom CMYK, however I would only use the input values and the reported values in Photoshop as a very loose reference point – I think that it is the actual results of the separations created that should be the judging criteria.


Stephen Marsh

Thanks for the reply.

The processes we are using can vary so much that it’s extremely hard to nail it down to a consistent standard.

The chemistry:
Although we mix / formulate our own cmyk inks from base pigments. Along with film production this is probably the most consistent part of the operation.

The screens;
This is the most variable part of the operation due to a multitude of reasons too many to list. It is however controlled as is practically possible.

The Material:
We always print on the same cloth. Although the white point of the cloth varies from batch to batch as dose the construction / absorbency, it falls within industry tolerances. These tolerances are far, far wider than any spec for paper would ever be.

Press
A press operator can expect to adjust a press for at least the first twenty minuets before it will stabilise as well as make adjustments during a print run. Lots of variables here that an experienced press operator can use to good advantage.

Targets and Instruments
We have had profiles made for us out of house (remote service) using the IT8.7-3 and i1 ECI targets. In house we use an SP68 (designed for reading textiles) print large targets (up to 18” x 36”) and measure the patches by hand - average of 10 readings per patch. Even how many threads there actually are per inch in the cloth where the patch is printed / measured is a variable and effects the light reflected from the sub straight under the cloth (I use a neutral grey)

With all the variables the only reason I think we can ever acheive any acceptable cmyk print is because our reproducible gamut, by flexo standards, is very limited, ie the fewer colours there are to reproduce the easier it is to reproduce them.

After 12 years of trying to nail a standard I’m beginning to think it is somewhat futile, the only reason for continuing the quest is of course professional pride - we can always do better. However when jobs fail even after remaking screens etc. I suspect it has more to do with image type, black generation and the black generation settings locked in the custom profile we are using.
My reasoning for using Photoshop was two fold, one to try and absorb all the variables. There are only 8 patches + a white to read. (A ‘phone enquiry to X-rite a number of years ago lead me to believe that fewer patches may, for us, produce better results). The second reason was for altering / experimenting with the black generation based on the image being prepared for print, ie is it dark or light, high contrast / low contrast etc.

So, to conclude do you still think I’m heading down the wrong road with Photoshop? Would a ColourMunki be better?
Other than give up, (which for me isn’t an option!) any other advice?

Thanks for you time.

Jon

 

[Lorna Wiles]

Thanks.

You input much appreciated.

Jon

 

[Stephen Marsh]

What sort of settings are you looking to use in Custom CMYK?

Black generation, black start, black limit, total ink limit (I understand that there will be multiple sets of settings to create a “suite” of profiles)?

There is a Fogra9 specification characterisation data for screen printing that is similar to Fogra39 in many respects similar in the gamut shape and size (slightly wider gamut in many areas).


Stephen Marsh

Thanks.

You input much appreciated.

Jon