Color Space Conversion

[sgordon]

We are in Packaging and are often challenged with weather or not to convert customer supplied Hi-res scans from their color space to ours. The color space we print in is GRACOL as we are G7. Most of our art/scans come in as SWOP. Converting the scan to GRACOL often makes a huge difference in how the 4/c is built. Are we making a bad move converting? Should we leave well enough alone or be doing/looking at something to help us decide if conversion is right or not?

 

[Tommyjt]

When you say SWOP I assume you mean U.S. Web Coated (SWOP) v2. Most of the images we receive come in that way as well and they have a very low level of GCR. If you want more GCR then it's a good thing to do. We use Alwan ColorHub to transform to GRACoL with a custom GCR setup. We get much more consistant press runs that way. I actually made my own ICC Profile using the GRACoL characterization data set because the GRACoL profile didn't have the level of GCR that I wanted.

 

[David Milisock]

sgordon I would assign the GRACOL color space to the swop files.
This will affect how the ​images display but will not affect the CMYK numbers sent to the process

 

[Stephen Marsh]

We are in Packaging and are often challenged with weather or not to convert customer supplied Hi-res scans from their color space to ours. The color space we print in is GRACOL as we are G7.

GRACoL is a specification. G7 is a calibration method that may be used as part of a larger method to achieve a specification such as GRACoL. So in my mind, being “G7” does not automatically/necessarily mean being “GRACoL”.

Most of our art/scans come in as SWOP. Converting the scan to GRACOL often makes a huge difference in how the 4/c is built. Are we making a bad move converting? Should we leave well enough alone or be doing/looking at something to help us decide if conversion is right or not?

If you are converting, you need to be able to trust the source profile is correct and not garbage. ICC profiles should provide certainty, however they do not. So you have two main choices:

1) Convert from the source to your destination and provide a proof
2) Assign/proof the files numbers/values without conversion in your destination and provide a proof

You may need to do both and ask them to pick!

Of course, the answer from one customer is not always going to equal the answer from another customer.

If there was no option to check/proof with the client, I personally would do #2 as David suggested – which is the traditional workflow of ignoring colour profiles and using the files as is. There is no magic bullet or easy answer to this decades old concern.


Stephen Marsh

 

[David Milisock]

Stephen Marsh, you know better than that, there is a magic bullet, unfortunately it's always aimed at our heads. An unskilled but self important client can be our worst enemy.

 

[Bill W]

If you convert 100% cyan from a U.S. Web Coated (SWOP) v2 color space to a GRACoL 2006 color space your 100% cyan will now contain 94% cyan, 0% magenta, 2% yellow and 0% black. The lab values will go from L 62, a -44, b -50 to L 59, a -39, b -47, which is, under CIE 1976, 5.38 dE off the Lab values for 100% cyan under ISO 12647-2 (2006) and above the dE tolerance of 5.

With a curve you will be able to print 94% cyan but even with digital plates you will be hard pressed to print 2% using conventional screening. With HD, or Maxtone or a hybrid screening you might be able to print it but you will need a very tight workflow to do so.

I agree with David and Stephan to use option 2 in Stephan's post. If you make sure your ink achieves the specs in ISO 12647-2 (2006) and they are gray balanced this option will serve you well, even with a bit more TVI in the highlight that one would find in an offset job that adheres to the same ISO 12647-2 (2006) spec and is gray balanced.

 

[Stephen Marsh]

If you convert 100% cyan from a U.S. Web Coated (SWOP) v2 color space to a GRACoL 2006 color space your 100% cyan will now contain 94% cyan, 0% magenta, 2% yellow and 0% black. The lab values will go from L 62, a -44, b -50 to L 59, a -39, b -47, which is, under CIE 1976, 5.38 dE off the Lab values for 100% cyan under ISO 12647-2 (2006) and above the dE tolerance of 5.

This is the case for pure solids, however when we add small values of other colours into the mix, the difference between original and conversion is greatly reduced. Apart from control elements such as colour bars, not many images have a pure solid primary as the major focus of the image. So in day to day production this is generally not a big issue. One can of course use CMYK to CMYK devicelinks in Photoshop, or one can create a simple action that preserves solid primary colour values from the original image in the converted image.


Stephen Marsh