How to use the Gretag MacBeth digital color checker

[Asher Kelman]

This is to kick of the discussion here, but I really want to follow the answers!

Well, we have purchased them! GM Color Checker Cards.

But how does one integrate the GM color checker into ones work RAW or jpg flow. I include jpg because some Panasonics are slow in writing to RAW and also because some busy event photographers do use jpgs routinely.

Using a white/black/grey card is pretty straightforward in work flow.

However, having done that, what does one do with all the other colors of the GM card?

What software exists to use this data or else templates to overlay on a card included in shoot, to get a more accurate color representation.

Asher

[scott kirkpatrick]

It's basically a simple profiling tool, and should be photographed in the light that you want to understand better, and processed using the tools (raw developer, RIP or print path) that you want to use for other real work. Then you can experiment to learn how to get its colors either right by eye or correct by measuring the rgb "eye-dropper" values in its patches. Michael Tapes can say much more in a few hours, since it is one of the first things he looks at when building a profile suite for RSP's or ETC's use. For personal use, it is a handy reference -- shoot it frequently and compare with what it looked like with another camera, or using another set of tools. Less frequently used than the WhiBal combination of white point, black point and white balance targets, but still useful.

[Asher Kelman]

Originally Posted by scott kirkpatrick

It's basically a simple profiling tool, and should be photographed in the light that you want to understand better, and processed using the tools (raw developer, RIP or print path) that you want to use for other real work. Then you can experiment to learn how to get its colors either right by eye or correct by measuring the rgb "eye-dropper" values in its patches. Michael Tapes can say much more in a few hours, since it is one of the first things he looks at when building a profile suite for RSP's or ETC's use. For personal use, it is a handy reference -- shoot it frequently and compare with what it looked like with another camera, or using another set of tools. Less frequently used than the WhiBal combination of white point, black point and white balance targets, but still useful.

Scott,

I religiously use both this card and a black white greay standard, although am currently testing the Whi-Bal™. However, I would like a simple program that can help me imrove on just getting the white balance right.

There must be software for the Mac to do so.

Is there?

Asher

[Michael Tapes]

Asher,

There is no trick to getting the White balance right. And you do not need a Color Checker to do it, but it will work fine.

For profiling the CC can serve as a reference, but it is not appropriate to create camera profiles from just the data of the CC. Gretag has their SG target for this, and the well known Q60 is a standard reference that was originally used for scanner profiles.

Back to White Balance. Whether you use a WhiBal or the Light gray patch of the CC, or any other certified reference, the process is the same. Shoot a reference shot in the same light as your subject, and set the WB in the RAW converter using the White Balance eyedropper. Then this Scene White Balance CWB (as I call it) can be applied to all of the other images that were shot in the same light in a few clicks. So you need nothing more than a certified WB reference, and your RAW converter to get it right. Obviously for mixed lighting situation, it can only be "right" for one of the lighting circumstances. You can either shoot the reference right at the subject of interest if there is one, or use an integrating tool such as the ExpoDisc to set an "average" WB for the scene.

Regarding the Gretag White/Gray/Black card, it is not useful for White Balance because the Gray is too dark. You could ramp up the exposure, but the chance of forgetting makes it a bad tool for the job. 18% gray is not good for WB unless overexposed, because there is simply less data down in at that luminance level then there is at light gray. White is not good because there is a large chance of clipping a channel which invalidates the WB reference if this occurs. Also with the exception of the Babel Target, White is very hard to get neutral. So IMHO a certified neutral reference of light gray is the way to go.

In discussing this with Thomas Knoll (father Photoshop and Creator of Adobe camera Raw), he stressed the point of not using White as a White balance reference. In fact he has coded it so that above a certain point ACR beeps and will not accept a high Luminance value as a WB reference.

Michael Jonnson, father of Capture One and Creator of RawShooter, is also in agreement, as is Magne Nilsen of ETC Consulting, who is a leader in the field Digital Camera profiles.

The only person I have spoken with who does not prefer light gray as opposed to White is Eric Hyman, father and author of Bibble.

[Christen Hansen]

The only person I have spoken with who does not prefer light gray as opposed to White is Eric Hyman, father and author of Bibble

.

Michael, please, what is your comment to Andy Rouse. In his Rawshooter Workflow he tells us just to put our cameras in Auto WB.

Then Kodak will, sorry, Rawshooter will do the rest.

Kindly Christen Hansen.

[Josh Liechty]

I was planning to purchase a Color Checker and use this ACR calibration script, until reading Michael's statements on camera profiling. By the "SG" target, do you mean the semi-gloss Color Checker with a greater variety of color samples? What kind of software would I use to calibrate with if I bought one of those? I notice that Gretag Macbeth's Eye One Match which I use with my Display 2 colorimeter has an option for camera profiling, but it is always grayed out.

[Andrew Rodney]

The ACR script requires the older 24 patch Macbeth. It will not work with the SG which is really designed for building true ICC profiles with some profile building application. So you need to figure out what you want to accomplish. You can't really build profiles for use in ACR. The "profiles" (which are not ICC profiles" are proprietary and built by Thomas Knoll. The script is kind of, sort of like a profile tweaker. So if you're using ACR, get the 24 patch target and the script and off you go.

The reason you can't use Eye-One Match is you only have the Colorimeter which unlocks just that one module. If you upgraded to use the camera profiler you could use the SG (but again, not in ACR).

[Asher Kelman]

Originally Posted by Andrew Rodney

The ACR script requires the older 24 patch Macbeth. It will not work with the SG which is really designed for building true ICC profiles with some profile building application. So you need to figure out what you want to accomplish. You can't really build profiles for use in ACR. The "profiles" (which are not ICC profiles" are proprietary and built by Thomas Knoll. The script is kind of, sort of like a profile tweaker. So if you're using ACR, get the 24 patch target and the script and off you go.

The reason you can't use Eye-One Match is you only have the Colorimeter which unlocks just that one module. If you upgraded to use the camera profiler you could use the SG (but again, not in ACR).

Andrew, this is the start of what I was looking for. So essentially, unless one wants a pseudo profile, ie not ICC, then the standard card is of little more than an expensive grey card!

Asher

[Andrew Rodney]

You can use the 24 patch card to build profiles in some packages but it's far from ideal (for that matter no target is). The 24 patch target was never made for anything other than a reference and for that it's decent.

[scott kirkpatrick]

Originally Posted by Andrew Rodney

You can use the 24 patch card to build profiles in some packages but it's far from ideal (for that matter no target is). The 24 patch target was never made for anything other than a reference and for that it's decent.

Since you're somewhat unaffiliated, can I ask your opinion on when and for what sort of profiling it is important to get a close match to targets, and when creative differences will leave most of us happier. I've used Imatest and the ColorChecker to see what C1 and RSP's profiles (both default and ETC) are doing, and all of them provide extra saturation, and some additional shifts seen in the stronger colors, when compared to standard targets. I also find the result, when applied to pictures, quite satisfactory.

scott

[Andrew Rodney]

I'm not sure what you're asking. Today's solutions treat digital cameras like scanners which isn't a very effective way to profile such a device. If we had the spectral characteristics of the chip and the illuminant per scene, we could undoubtably build better profiles (and in theory, on the fly per image). We're not there yet but in a year, maybe.

The SG is the "best" general purpose target today for building camera profiles. Its a compromise over the DC which was built to build camera profiles (unlike the 24 patch chart). The DC had glossy patches. Good for extending the gamut of the target and therefore the resulting profile (remember, digital cameras don't really have a gamut but rather a color mixing function based on what you put in front of them). The glossy patches were VERY difficult to deal with unless you had the scene lit perfectly as reflections hosed the profile. The SG is a Semi Gloss (hence the name).

I don't spend much time anymore building camera profiles because the RAW converters I use don't benefit from them (or can't use them) and the results are almost always iffy. And with the better product I have to use (ProfileMaker Pro), there are so many possible options in building a profile, I'd have to be compelled to send a good day or three doing this and testing dozens of profiles.

If I were doing studio work, or copy work, that would be a useful endeavor. But for just about anything else, its far too much work for the questionable benefits.

[Michael Tapes]

I think all of Rodney's point, point to why many of us let Magne Nilsen build our camera profiles (for Capture One and RawShooter, at least). He seems to be the master of the camera profile, but even for him it is a several week (part time) effort to produce the final profile for a given camera models (from many "profile suites from many cameras that I usually provide).

Really good camera profiles are not something that one can just generate from shooting a target and running it through some profiling software. It is a very scientific and aesthetic endeavor, and in the end requires, real world images to do the final tweaking.

As Andrew points out, even for an expert like himself, this is not a simple 1 hour process.

[Michael A. Landré]

"but it is not appropriate to create camera profiles from just the data of the CC"

Andy and Michael:

Not meaning to be impolite, but I think that you should mention
that it is possible to use the Macbeth Color Checker and
this ACR calibration script to generate data for "calibrating"
the ACR Camera Profile when using Adobe CS2 for raw conversion.

This is what I - and many others have done and I am very happy
with the results for my Canon 20D and 5D.

For those who use CS2 for raw conversion, the Macbeth Color Checker and the ACR calibration script are a very useful tool.

[Asher Kelman]

Sorry, I don't follow. Which script are you referring to?

Asher

Originally Posted by Michael A. Landré

"but it is not appropriate to create camera profiles from just the data of the CC"

Andy and Michael:

..snip..but I think that you should mention
that it is possible to use the Macbeth Color Checker and
this ACR calibration script to generate data for "calibrating"
the ACR Camera Profile when using Adobe CS2 for raw conversion.

[Michael A. Landré]

this ACR calibration script: http://fors.net/chromoholics/download/

The script/link refered to in Josh Liechty`s post on Page 1

[Harvey Moore]

I'll second Michaels recommendation for the chromoholics script for acr camera calibration.

I too use it with my 20D & 5D

harvey

[Michael Tapes]

Originally Posted by Michael A. Landré

"but it is not appropriate to create camera profiles from just the data of the CC"

Andy and Michael:

Not meaning to be impolite, but I think that you should mention
that it is possible to use the Macbeth Color Checker and
this ACR calibration script to generate data for "calibrating"
the ACR Camera Profile when using Adobe CS2 for raw conversion.

Yes, you are right, but that does not create a camera profile, and not all are pleased with that process. But that does not mean that it is not valid for many or even most users of the script.

The CC simply does not offer the "data" to support the creation of a sophistocated camera profile from scratch, as the only data. Since Camera Raw is based on 2 target shots for each camera, the internal "profile" already is built and the calibration script is a modification upon that.

So I do not think we disagree on the CC, just a matter of the semantics of "camera profile".

[Andrew Rodney]

Andy and Michael:

Not meaning to be impolite, but I think that you should mention
that it is possible to use the Macbeth Color Checker and
this ACR calibration script to generate data for "calibrating"
the ACR Camera Profile when using Adobe CS2 for raw conversion.

I mentioned it in my first post...

The site can be found here:

http://www.fors.net/chromoholics/download/

The current script needs to be updated for ACR 3.4 however.

[Asher Kelman]

The new Beta 3.8 is now available for download and works with ACR 3.4

Asher

Originally Posted by Andrew Rodney

I mentioned it in my first post...

The site can be found here:

http://www.fors.net/chromoholics/download/

The current script needs to be updated for ACR 3.4 however.

[Herb]

On the question of best tone (white, light grey, dark grey) for White Balance, surely all views can be reconciled?

WB of a raw file is a single point correction. Based on my vague recollection of school maths, the best spot on a characteristic to make a single point correction is dead centre.

A keen amateur photographer who is aware of the need to 'expose right' will place white at the max point. A grey one stop down will then be centre of the characteristic. Proves Michael Tapes and similar are right.

However, sloppy professionals who snap away on auto everything know that clipping whites on a wedding dress is death. They habitually underexpose by one stop 'just to be safe'. White for them will sit dead centre of the raw characteristic and will be just the point to do the WB. Proves that for the vast majority of photographers, Eric is probably right.

Of course, since we are dealing with nice linear sensors, you could even expose a dark grey target (or any other shade) so that it gets placed centre of the characteristic and use it for WB.

The interesting thing is that lots of folk are not convinced that everything does work perfectly linearly, so all of this, including 'expose right' becomes questionable.

Life was simple before John Sheehy.