question about color tolerancing

[gordo]

Is there any ISO/GRACoL 7 specification for color tolerances for anything other that printed solids?
My understanding is that a printed solid must be less than a Delta E of 5 (1976?)
But what about an arbitrary color build? E.g. The patches in an iT8 target? Must they also be less than a Delta E of 5?

 

[Stephen Marsh]

Gordo, I do not believe that ISO de76/ab tolerances are restricted to only solids (at least for proofing, however I am sure that press side has the same tolerances).

Solid primaries are indeed <5 dE76. Solids also have a dH tolerance of <2.5 which is a balancing act between acceptable dE and dH.

The maximum deviation for key screen tint builds that are not grey balance is <6 de76. For example, the Fogra media wedge contains patches such as: 70c70m70y, 40c100y, 100c40m, 70m60k, 40m40y, 10c40m40y.

There are also grey balance patches, such as 10c6m6y, 20c12m12y, 40c27m27y, 60c45m45y, 80c65m65y - these have an average hue tolerance of <1.5 dH.


Stephen Marsh

 

[gordo]

Gordo, I do not believe that ISO de76/ab tolerances are restricted to only solids (at least for proofing, however I am sure that press side has the same tolerances).

Solid primaries are indeed <5 dE76. Solids also have a dH tolerance of <2.5 which is a balancing act between acceptable dE and dH.

The maximum deviation for key screen tint builds that are not grey balance is <6 de76. For example, the Fogra media wedge contains patches such as: 70c70m70y, 40c100y, 100c40m, 70m60k, 40m40y, 10c40m40y.

There are also grey balance patches, such as 10c6m6y, 20c12m12y, 40c27m27y, 60c45m45y, 80c65m65y - these have an average hue tolerance of <1.5 dH.


Stephen Marsh

Thanks Stephen, that's very helpful. Every reference data sets that I have seen appear to only provide the target value but do not associate a tolerance pass/fail to any of the patches. Indeed devices like the Techkon SpectroDENS let you set whatever tolerance values you want.
Is there an official downloadable document that has the information as to what the pass/fail tolerance values are for the iT8 chart (or similar)?
Or is it up to the vendor to decide what the tolerance is for specific patches?
Or is it up to the individual print shop to decide on their own tolerance values?

Thanks.

 

[Stephen Marsh]

I’ll probably need to look into this deeper Gordo, I too hope that others can shed more light.

My day to day proofing world is ruled by the 72 patches on the Fogra media wedge, which is a lot less than full characterisation chart.


Stephen Marsh

 

[gordo]

I’ll probably need to look into this deeper Gordo, I too hope that others can shed more light.

My day to day proofing world is ruled by the 72 patches on the Fogra media wedge, which is a lot less than full characterisation chart.


Stephen Marsh

Thanks! That would be great! I can't understand why a standards/specifications group or instrument vendor wouldn't automatically include the appropriate tolerances whenever the publish a specification so that folks wouldn't have to dig deep.
If any other color guru can point me in the right direction it would be appreciated.
Thanks again.

 

[Stephen Marsh]

Gordo, try page 7 of the following PDF:

20090625_akraushaar_proofing_psrv4_en.pdf

eci.org/_media/downloads/presentations/20090625_akraushaar_proofing_psrv4_en.pdf

Stephen Marsh

 

[gordo]

Gordo, try page 7 of the following PDF:

20090625_akraushaar_proofing_psrv4_en.pdf

eci.org/_media/downloads/presentations/20090625_akraushaar_proofing_psrv4_en.pdf

Stephen Marsh

Thanks again!

So, what I get from this is that there are specified tolerances based on CIE Lab 1976 for:
1 the paper (Delta E 76) DE 3
2 the primaries (Delta E 76) DE 5 - CMYK but not the secondaries RGB. Or do they consider RGB as primaries?
3 the primaries (Delta H) - CMYK but not the secondaries RGB is less than or equal to 2.5.
4 that grey balance is measured as Delta H not E and is less than or equal to 1.5
5 that the Delta E of all colors is less than or equal to 3
6 "Delta E Max of 6" is undefined. Is it the paper? I.e. It should be a Delta E of 3 but 6 is OK?

Is this correct?

 

[Stephen Marsh]

Yes, Gordo, that basically covers it… I’ll try to dig up a little more info on a couple of outlying points…

Basically, everything <6 dE (all 1, 2 or 3 colour tint/solid builds), unless a solid 1C primary, which is <5 dE and <2.5 dH, unless a grey balance patch, which is < 1.5 dH. Average tolerance for all patches should be < 3 dE. Maximum tolerance for any single patch should be < 6 dE. Substrate should be <3 dE.

You will note that on the previous photo attachment of certification labels from both Kodak and CGS products, that both use the same values, these are not vendor created standards – they are both conforming to the same specification. The target values from the specification are all M0 measurements and should be verified with a unfiltered (no UV cut) spectro. Also note that all of my examples were biased toward proofing for the current Fogra39 / ISO Coated v2 ISO 12647-2 conditions (GRACoL and SWOP may have different requirements).


Stephen Marsh

 

[Stephen Marsh]

Over and above what is required for passing a colour bar, there may be some other metrics used by other QA software, perhaps with full characterisation charts (and not just a limited amount of patches from a colour bar). In the case of X-Rite’s i1profiler Quality tool:

The All patches result is the average dE of every color patche in the reference color set.
The Lowest 90% result throws out the highest (worst) 10% of the dE values, then calculates the average dE.
This value will therefore be lower (better) than the All patches result.
The Highest 10% result throws out the lowest (best) 90% of the dE values, then calculates the average dE. This value tells you the matching quality of your most difficult colors.
This value will therefore be higher (worse) than theAll patches result.

Stephen Marsh

 

[gordo]

Super thank you and very helpful. It still amazes me how difficult it is to get what one would think is basic information.