New ColorChecker Passport vs. Old ColorChecker Classic?

[Redcrown]

I got a new Xrite ColorChecker Passport. Shot it side-by-side with my 10+ year old ColorChecker Classic. Daylight, highly diffused dull overcast day. Very even lighting. ACR White Balance came out about 6250/+25. Plus or minus 100.

Both charts generate the same WB, but there are differences in the color patches. Faint differences, but easily visible on screen.

In the dcamprof documentation, Torger says Xrite changed the colors in November, 2014, so I assumed that explained the differences I'm seeing. But after some Googling, I can't find any confirmation of that. In fact, I've only found only 4 sources for the numerical reference values of the CC24. They all match, but they are all dated prior to 2014.

I sent an e-mail to Xrite over 10 days ago asking where I can find new reference charts. No reply. Xrite is one of the sources I found, but the only chart they show is dated 1/22/2010, updated 4/20/2012.

So, can anyone show confirmation that a new passport is really different than an old classic? Anybody know where there are post-2014 reference charts (LAB values)?

[Schewe]

Well ColorCheckers are known to fade over time. I wouldn't be inclined to trust a 10yr old chart...but that's just me.

[Lundberg02]

Where was it stored?  Did it look faded? Probably should trash it.

[Redcrown]

My old CC24 has been stored in an opaque envelope in a drawer. Used very little, only for profiling. It might still be faded, but that's the question. Should an old CC24 Classic match (reasonably) a new Xrite Passport? Or was the Passport changed in 2014?

All the LAB reference values I have are pre-2014. Are there new LAB reference values for a new Passport?

[AlterEgo]

In the dcamprof documentation, Torger says Xrite changed the colors in November, 2014

ColorChecker Classic and ColorChecker SG... not Passport

and then do buy spectrometer... for example EFI branded i1Pro shall be quite not expensive... that will address the issue.

[AlterEgo]

Well ColorCheckers are known to fade over time. I wouldn't be inclined to trust a 10yr old chart...but that's just me.

you just need to use your spectrometer (I bet you have more than one) and software that allows you to use measurements (hint - no, not Adobe DNG PE and not X-Rite OEM one)

[Doug Gray]

you just need to use your spectrometer (I bet you have more than one) and software that allows you to use measurements (hint - no, not Adobe DNG PE and not X-Rite OEM one)

Right, that's the way to go.  I have 4 or 5 ColorCheckers some dating back 10+ years.  I'm actually surprised how little they fade. I've found more variation between Colorchecker vintages than the relatively little they change over time. So, even with new ones, I find it useful to measure them.

[Redcrown]

There is some interesting logic here. The ColorCheckers are expensive pieces of paper and plastic. The expense being justified by the claim they are manufactured to exacting standards and guaranteed to be accurate within acceptable tolerances.

So why should you need to buy an equally expensive spectrometer to verify that your ColorChecker is any good? And why should you trust that the spectrometer produces more accurate readings? What do I need to buy to check my spectrometer?

For those that have checked the ColorChecker, did you find significant differences? What was the Delta-E? 

[Doug Gray]

There is some interesting logic here. The ColorCheckers are expensive pieces of paper and plastic. The expense being justified by the claim they are manufactured to exacting standards and guaranteed to be accurate within acceptable tolerances.

So why should you need to buy an equally expensive spectrometer to verify that your ColorChecker is any good? And why should you trust that the spectrometer produces more accurate readings? What do I need to buy to check my spectrometer?

For those that have checked the ColorChecker, did you find significant differences? What was the Delta-E?

dE76's as much as 3.5 but only in highly saturated colors where perceptual differences are lower. dE2k variation only just over 1 max. The neutral patches, which have similar sensitivity to dE76 and dE2k, are quite close between the colorcheckers and have remained unchanged over time. However, there is significant variation in the white patch amongst them.  I always use the second whitest patch for white balance because of that.

I didn't buy spectros to check colorcheckers but to profile printers and monitors. I have several generations of I1 Pros and they are much closer to each other than the colorcheckers.

[bjanes]

So, can anyone show confirmation that a new passport is really different than an old classic? Anybody know where there are post-2014 reference charts (LAB values)?

Check this this

Bill

[AlterEgo]

why should you need to buy an equally expensive

to get from "90%" to "95%" of something you are really interested in... now to get from "95%" to "97%" you spend "$10K+"... from "97%" to "98.*%" you spend "$100K+"

some people do have perfect mental happiness with the end results (photos) building profiles off unmeasured targets with software that does not know the difference between passport, colorchecker classic or just subset of patches from colorchecker SG, etc, etc and does not allow your to enter your own measurements (Adobe DNG PE for example)... and some people have a bug that does not allow them to be happy... your choice where you belong

[fdisilvestro]

... and some people have a bug that does not allow them to be happy...

  how true!

[Redcrown]

I asked Xrite if they guarantee their ColorChecker targets to be within some measurable level of tolerance. They actually answered, but dodged the question. Sounds like they guarantee the product until it leaves their facility. Once you "handle" it, the warranty is void.
---------------------------

My inquiry:

In a Luminous Landscape forum discussion, some people claim you need to use a spectrometer to read an Xrite ColorChecker target in order to get accurate values. Some who have done that claim a 10% difference (error) between the actual color patch values and the published values.

What do you think of that? I can not find any statement of guarantee on your website for the accuracy of ColorChecker targets. Is a 10% variance within your manufacturing standards?
---------------------------
Their Response:

I'm sure you can understand that there are many variables when it comes to measuring a ColorChecker target that we've had in production for over 40 years now.

The age of the measured target is the first variable, and we would strongly suggest that you work with a ColorChecker target that is less than two years old.

The second variable is the type and calibration state of the spectrophotometer used to make measurements.

While Luminous Landscape is a well respected imaging website, forum discussions can include incomplete, or even inaccurate comments.

We will guarantee the quality of any newly purchased ColorChecker target, but there will always be some degree of variability in any manufactured color target. People who wish to use the ColorChecker target in X-Rite's i1Profiler software can either use the standard values that we provide in our software, or they could choose to measure their own specific ColorChecker target, if they felt this was going to give them a more precise result.

Please understand that once a ColorChecker target leaves our facility, we have no method to determine whether a target was handled correctly. I've seen many online comments coming from people who were using a ColorChecker target that would be old enough to vote, and no reliable information on how it had been stored or handled.


Thank you, 

Bruce Wright
Technical Support Rep - Level 2
X-Rite Incorporated
--------------------------
P.S. I'm an old programmer. Debugging makes me happy. I love chasing those suckers down and driving a stake through their heart. The hunt is usually more satisfying than the kill.

[Doug Gray]

FWIW, I have a 2003 CC and 2012 CC. Measured now with PatchTool, the bluer patches are slightly faded on the older one and the dE2k of the worst patches are 1.5 and 1.3 respectively. Average dE2k is .6

[GWGill]

There is some interesting logic here. The ColorCheckers are expensive pieces of paper and plastic. The expense being justified by the claim they are manufactured to exacting standards and guaranteed to be accurate within acceptable tolerances.

You are assuming the expense is justified by that - I'm not sure they actually claim any such a thing.

Personally I assume the expense is simply for the convenience of being able to purchase a ready made test chart, manufactured with uniformity and stability as an aim, and composed of spectrally diverse samples. If I'm at all interested in high color accuracy, then the only safe thing is to measure the chart myself.

So why should you need to buy an equally expensive spectrometer to verify that your ColorChecker is any good?

Because you are interested in the color accuracy of your whole workflow, and don't wish to rely on the vagaries of the chart manufacturing process ?

And why should you trust that the spectrometer produces more accurate readings?

Because the instrument is actually calibrated against a reference at the factory, rather than relying on the reproducibility of a manufacturing process ?

What do I need to buy to check my spectrometer?

An even more expensive instrument ? 

But more seriously, reflective calibration of the instrument is done against a white reference that is specifically manufactured for longevity, and individually calibrated. Something like the i1pro2 also has a wavelength reference to calibrate against.

[howardm]

What is the best/correct way using (I assume) spotread to read the CC (assuming there is some sensitivity to the command options)?

[AlterEgo]

What is the best/correct way using (I assume) spotread to read the CC (assuming there is some sensitivity to the command options)?

whatever you do do not forget to to heat the spectrometer's lamp ("restore the lamp") otherwise you will get a drift (may or may not be significant for your purposes)...

[howardm]

The only way I can see doing that (since I dont own Patchtool) is maybe via i1Diagnostics.  It seems like spotread doesn't support it in the current 1.8.3 version.

Different question:  Can I use CT&A to read spot colors?  I'm considering CT&A and I'd prefer not to buy Patchtool if I didn't have to.

[AlterEgo]

The only way I can see doing that (since I dont own Patchtool) is maybe via i1Diagnostics.  It seems like spotread doesn't support it in the current 1.8.3 version.

didn't GWGill add the functionality to his relevant Argyll utility ? may be still in beta ...

[howardm]

I saw his thread(s) on the colorsync list etc etc and it seemed like he talked about it in the Feb 2016 timeframe but the last release was back in late 2015 so yea, I'm thinking 'still in beta'