Colorchecker Passport Help

I just received my x-rite Colorchecker Passport and could use some help creating custom DNG profiles. Here's what I've done:

1. Set a custom white balance for the scene. Not a requirement for making a custom profile according to x-rite, but doesn't hurt either.

2. Shot an image outdoors in sunlight (no clouds) using sunny 16 (f/16, ISO 200, 1/200s) to ensure good exposure. The histogram for the resulting image is decent.

3. Brought the RAW file into ACR and created a DNG.

4. Dragged the DNG into the window on the Colorchecker Passport software and created the custom profile.

The problem is that when I go back into ACR with an image of the Colorchecker Passport and try different profiles on the calibration tab, Adobe's "Camera Neutral" profile seems to match some of the actual color swatches better than the new profile created using the x-rite software.

I then downloaded the Adobe DNG Profile Editor (beta 3) and created a custom profile using that software using the same DNG image used for the x-rite software. The result is a bit different and appears to be slightly closer for some of the actual color swatches than the one created by the x-rite software, but off on others. Following are my results. I'd really appreciate knowing from the color gurus here which one looks to be the closest match to the Gretag Macbeth color swatches. To me none of them do a great job.

Using profile created by Colorchecker Passport software:

And here's the same image using "Adobe Standard"

Well ... one difference solved. When I created the profile using the Adobe DNG Profile Editor I had "Both Color Tables" (2850 and 6500K) selected. When I did it again and selected "6500 K only" the resulting profile matches the one created by the Colorchecker Passport software for every swatch. Anyone know when you would use the "Both Color Tables" option?

The both color tables option is for dual-illuminant profiles.

There are two ways to make camera profiles using the Adobe DNG profiler software.
One is to create a profile using a single test image. If you choose the software will then extrapolate the readings for the other luminate and create a duel luminate profile.

The other, more accurate approach is to shoot the color checker in both light sources, then import both files and use them to create a duel luminate profile. The differences are subtle but important for good color management.

Once again illustrating the fact that the nut holding the calibrator simply needed to be adjusted properly!

Its also important to remember that the color checker and profiling is useful for customizing the white balance and color response for a specific light source in a specific situation.

We shooting film, it was common to buy a several boxes of film from the same emulsion batch, then shoot tests and have them processed in a color processor with a known chemical age. Then build a color correction filter pack to shoot a color sensitive product. Naturally, using the filter pack, film from the same emulsion batch and processed in the same chemical batch.

Today, put a color checker in the light source that will be used for the shoot. (Remember to do it after the power settings are finalized. Then shoot the color checker and do a profile to be used for that shoot.

It makes color managed workflow that much more accurate. I do this for artwork when shooting for an exhibition catalog or gallery owner.

Wilt wrote in post #14967967
Once again illustrating the fact that the nut holding the calibrator simply needed to be adjusted properly!

LOL!!!! Thanks for your wisdom

Thanks Chris and David.

So, does the image of the color swatches produced by the Colorchecker software look the best to you?

Bob_A wrote in post #14968568
Thanks Chris and David.

So, does the image of the color swatches produced by the Colorchecker software look the best to you?

Personally, I don't judge color rendition based on the swatches, but rather the effect the generated profile has on real-world images.

But I generally prefer the Colorchecker profiles.

Thanks again Chris.

The Colorchecker profile gives a pretty different result. Blue sky images with my D700 are typically a bit magenta using "Camera Neutral" which changes to leaning towards cyan with the Colorchecker profile. After getting used to the Camera neutral rendition the change is going to take awhile to get used to.

chris_holtmeier wrote in post #14968625
Personally, I don't judge color rendition based on the swatches, but rather the effect the generated profile has on real-world images.

But I generally prefer the Colorchecker profiles.

The color swatches on the color check are irrelevant.
What is important is how the profile adjusts the native color spectrum of the sensor and adjusts it to the standard reference for the color checker. i.e. the Passport itself in the lightsource used to capture the image. The color checker is 24 color squares that are used to create an ICC profile (or proprietary Adobe profile).

My experience is that its almost impossible to get a color checker image on a computer screen to look exactly like a color checker held next to the screen.

What is critical, is being able to create an image, that when exported from Lightroom or Photoshop and printed looks nearly identical to the artwork that was photographed.

That can be accomplished using profiles as described.

dmward wrote in post #14969770
The color swatches on the color check are irrelevant.
What is important is how the profile adjusts the native color spectrum of the sensor and adjusts it to the standard reference for the color checker. i.e. the Passport itself in the lightsource used to capture the image. The color checker is 24 color squares that are used to create an ICC profile (or proprietary Adobe profile).

My experience is that its almost impossible to get a color checker image on a computer screen to look exactly like a color checker held next to the screen.

What is critical, is being able to create an image, that when exported from Lightroom or Photoshop and printed looks nearly identical to the artwork that was photographed.

That can be accomplished using profiles as described.

Well, I just reprocessed a bunch of vacation images using a "daylight" profile that I made with the Colorchecker and it does look like they are significantly better than using "Camera Neutral" or "Camera Standard". So far I'm sold on the device/software

The same profile created using the Adobe DNG Profile Editor software yields identical results.

So, what if the "artwork" is the CC card? Your argument makes little sense, and yet there is a reason that placing emphasis on the CC colors is not super important. Of all the things you would like to validate with a custom profile made from a CC card would be the ability of the custom profile to replicate the CC card color under the lighting conditions in which it was shot. Rarely do you know the Lab or RGB values of the colors in a scene like a landscape or a portrait - you do actually know the reference RGB or Lab values for things like the CC card, QP card, etc.

You can use these values to adjust the custom profile if the visual or color picker values rendered with the custom profile just don't quite cut the mustard. Of course, if accuracy is not the main point of creating a custom profile, you can always eyeball it - but the eye (brain) is notorious for making things seem right, even if they are not quite right and vice versa.

If your display is not properly calibrated and profiled, you can use any custom camera profile you want and you still will not get a visual match to the lit target. Traditionally, and over-simplistically, you need to use a fully color managed workflow to attempt a match. This includes:

- controlled, constant, reproducible lighting to shoot the target AND view the target for purposes of the above comparison;

- custom camera profile (DNG or ICC)

- Calibrated and profiled display, with target settings that match the viewing conditions under which the target will be illuminated for shooting and comparison purposes.

This way all of the devices will be playing by the same rules and you can try to make a meaningful comparison and adjust the chain of devices as necessary. The "devices" in this example are:

the lit target+ camera
the display
the lit target

While there are issues related to reflected versus transmitted light, the same principles apply here as when comparing a displayed versus printed image.

HOWEVER - getting a match to the CC or similar chart does not mean you will get pleasing or accurate digital camera images, due to the nature of acquiring the target image and the way that the CC data are treated. That is, sometimes custom profiles create more problems than they solve, colorwise. So, you may be able to make pretty pictures of the target, but your normal photographic images may not fare so well. THere is a very interesting discussion of some of these concepts here:

http://www.c-f-systems.com …ColorIntegrityC​FS-276.pdf

So, while the patches are relevant, trying to recreate them exactly may topple the whole house of cards.


kirk

dmward wrote in post #14969770
The color swatches on the color check are irrelevant.

...

My experience is that its almost impossible to get a color checker image on a computer screen to look exactly like a color checker held next to the screen.

What is critical, is being able to create an image, that when exported from Lightroom or Photoshop and printed looks nearly identical to the artwork that was photographed.

That can be accomplished using profiles as described.

Really interesting reading Kirk. Have you tried the C F Systems solution to see if it indeed does a better job?

Also, it would be interesting to know how Adobe, Nikon, Canon, etc. come up with the different profiles for their cameras. Are they created using the same type of device as the Colorchecker or do they use a different method?

Interesting stuff, which brings me back to my initial frame of mind before I started this thread. How do I know if what the Colorchecker is better than the profiles supplied in-camera or by ACR, DPP or Capture NX2? From what I've seen so far colors SEEM to be better than, for example, using Camera Neutral using a profile created under the same lighting, however after reading the pdf it would be good to know if the Colorchecker method is a good approach. The article concludes that: "With a camera profiled in this patchwork quilt manner colors can shift several times going from darker to brighter illumination - color integrity is not only lost but becomes nearly impossible to regain."

So ... is C F Solutions correct or are they spreading misinformation in order to sell their own competing solution?