View Full Version : Instruments limitations

12-14-2010, 05:26 PM
I am using instrument X to measure ColorFlow custom charts.

The problem with instrument X is that it does not correlate well with other instruments I have in my stable.

In particular, solid magenta, which is the biggest problem.

To give you an idea, while instrument X reports b* = -10, my handheld instruments report b* around -6. Do you see the discrepancy?

Does ColorFlow incorporate any resources for dealing with this thorny situation?

Best / Roger

12-14-2010, 06:32 PM
Hi Roger,
Yes, this is an overall problem that is very difficult to fully solve. The x-rite XRGA stuff comes close to implementing inter-instrument guidelines for alignment but does not support all equipment (for good reason). Not sure if this negatively impacts you, as the device types are not listed in your post. Nonetheless, a -4 delta b* is a pretty big discrepancy between devices. I must say that the XRGA standard is meant to accommodate for very small shifts between 0.5-1.5 dE due to different calibrations schemes.

Looking forward, it will be important for ColorFlow to support the XRGA implementation, as this will be a new calibration scheme offered by x-rite. Until we have more information on the new Device Services SDK that x-rite has created, we will continue to support the current Macbeth calibration scheme.


12-15-2010, 05:42 PM

I am keeping my fingers crossed that XRGA will help alleviate some of the inter-instrument agreements we all run into.

One thing, though, about ColorFlow's present metrology policy, with regards to filtering UV excitation or not off the measurements. With the EyeOnePros, it's easy to control ColorFlow behavior. One merely uses a UV-cut flavor and ColorFlow has no choice but to consume that mode of measurements. With the DTP70 (I'm not sure is supported?), it's easy to control what kind of spectra one feeds to ColorFlow. But using the i1Sis, the choice is not under the user control -- yet, I gather.

The results is that, the b* of -10 on the solid magenta ink that the iSis reports with no UV filtration, comes out as b* of -7 with UV filtration.

In the past, I tried to stay away from UV filtration like the pleague. On press, I realize there's little we can do with regards to the amount of FWA built into the printing substrates. But on the proofing papers, it's possible to select a paper that shows little or no fluorescence.

The problem is one of traceability, David. In the sense that, all the "standard" dataset or specifications that exist off the shelf for us, poor printers, to pick and use in our production, are all devoid of any UV filtration. IDEAlliance SWOP2006_C5 is a direct descendent of the 1995:TR-001 data which, as you know, comes from the averaging of X-Rite 938 and Gretag SPM-100 unfiltered spectral data. So, what is a poor printer to do confronted to ColorFlow's UV filtered data? How will I reconcile unfiltered Reference data to filtered data?

Part of the problem stems in the instrument itself and the choice of illumination. Right now, the LED lamps in the iSis don't always yield comparable results with a tungsten-based illumination. That's one big problem. That's why, I pray that XRGA will allow correcting for the lighting effect (but I know I should not hold my breath for this).

Until I started to use ColorFlow, when confronted with inter-instrument agreement problems, I use to edit the measurement data, using the tools I have at my disposal, BEFORE turning them into press profiles as a basis for calculating near-neutral curves. But, now, I'm completely at the mercy of ColorFlow :(

I realize in the present v1.1.1.14 incarnation, ColorFlow does not leave me the option, and whatever comes out of its calculation is whatever is going to be put on press.

Oh well, at least it's going to afford me with some starting point. But I'm going to be able to take the generated reports at face value anymore, because, as you know, we're not comparing apples with apples. Instead, when I want to compare my press calibrated data, the comparison will be made between Unfiltered Reference dataset and filtered dataset. That's unfortunate.

Best / Roger