[swanlefitte]

This article checks white cards and reports the differences. The variability is worth noting. If you find your card gives an unsatisfactory look you can use this to see what card you might like.
Leeming LUT Pro - White Balance Super-Test

[photoptimist]

It's also worth noting that if they re-ran this test on different sunny days that had different UV index levels (e.g., winter vs. summer or morning vs. noon), they'd get different results with some of the materials.

Ordinary "white" paper, "white" cotton clothing, as well as some laundry detergents contain optical brighteners that fluoresce -- they convert UV light into white or blueish light that helps make the object (which might have yellowish hue due to residual wood pulp color or dirt) look whiter. During high-UV index days and times, the material will be brighter and probably bluer. During low-UV index days and times, the material will be duller and probably yellower or redder.

Perhaps one good test for any prospective white-balance or gray card is to shine a black light on it. If the card glows, it's a bad choice.

[BigMackCam]

Very interesting - thanks for posting this. I wouldn't have expected such considerable differences.

Another thing worth noting - in addition to photoptimist's observations above - is that the accuracy of these references can change over time. I believe X-Rite recommends a new ColorChecker Passport every two years for this reason, and whilst that may be conservative and lean towards revenue generation rather than necessity, it's reasonable to assume that, say, a five year old reference chart might not be as accurate as a factory fresh example.

For us amateurs, I think it's more important to be able to achieve consistency in white balance rather than 100% accuracy, especially when shooting raw. That way, assuming consistent lighting conditions, a custom white balance taken from a reference shot will be largely valid across an entire group of photos, and one chosen set of corrections or tweaks can be applied to that group. Since white balance is typically giving us a "starting point" rather than an accurate representation of the actual lighting conditions we're shooting under, absolute accuracy doesn't strike me as crucial. For that reason, I'm still using the ColorChecker Passport I bought five years ago, both for colour profiling and (occasionally) white balance - and whilst the accuracy might not be 100% by now, it's good enough

[stevebrot]

Conspicuously missing would be the traditional Kodak gray/white cards (18%/90%).


Steve

[swanlefitte]

Conspicuously missing would be the traditional Kodak gray/white cards (18%/90%).


Steve

The Shirley Card?
How Kodak's Shirley Cards Set Photography's Skin-Tone Standard : NPR

[stevebrot]

The Shirley Card?
How Kodak's Shirley Cards Set Photography's Skin-Tone Standard : NPR

Nope...Kodak R-27

Kodak R-27 Gray Cards 1903061 B&H Photo Video


Steve

[photoptimist]

Very interesting - thanks for posting this. I wouldn't have expected such considerable differences.

Another thing worth noting - in addition to photoptimist's observations above - is that the accuracy of these references can change over time. I believe X-Rite recommends a new ColorChecker Passport every two years for this reason, and whilst that may be conservative and lean towards revenue generation rather than necessity, it's reasonable to assume that, say, a five year old reference chart might not be as accurate as a factory fresh example.

For us amateurs, I think it's more important to be able to achieve consistency in white balance rather than 100% accuracy, especially when shooting raw. That way, assuming consistent lighting conditions, a custom white balance taken from a reference shot will be largely valid across an entire group of photos, and one chosen set of corrections or tweaks can be applied to that group. Since white balance is typically giving us a "starting point" rather than an accurate representation of the actual lighting conditions we're shooting under, absolute accuracy doesn't strike me as crucial. For that reason, I'm still using the ColorChecker Passport I bought five years ago, both for colour profiling and (occasionally) white balance - and whilst the accuracy might not be 100% by now, it's good enough

Good points, Mike.

The X-Rite recommends a new ColorChecker Passport because some of the color chips probably fade or shift in color over time due to exposure to light, humidity (which enables reactions between the colors and chemicals in the paper or backing), ozone (which oxidizes chemicals), and heat (which accelerates all kinds of deterioration). You are no doubt right that the 2-year recommendation is conservative and probably only needed by professionals who create fashion catalogs or museum archival documentation.

In theory, if one had a second "good" ColorChecker (or a very stable light source), it could be used to calibrate an older, faded ColorChecker and regain accuracy.

P.S. It would be interesting to see what happens to a ColorChecker Passport left out in full sun or in front of a strong UV source. (placing a series of thin strips of aluminum foil on the ColorChecker to cover half of each tile and then leaving it in sunny south-facing window could be a fun experiment!)

[BigMackCam]

Good points, Mike.

The X-Rite recommends a new ColorChecker Passport because some of the color chips probably fade or shift in color over time due to exposure to light, humidity (which enables reactions between the colors and chemicals in the paper or backing), ozone (which oxidizes chemicals), and heat (which accelerates all kinds of deterioration). You are no doubt right that the 2-year recommendation is conservative and probably only needed by professionals who create fashion catalogs or museum archival documentation.

In theory, if one had a second "good" ColorChecker (or a very stable light source), it could be used to calibrate an older, faded ColorChecker and regain accuracy.

I seem to remember reading a (non-scientific) test on another forum where a chap used the color picker tool in his post-processing software to analyse differences between photos of his several-years-old but well-looked-after ColorChecker Passport and a brand new one. His tests showed that there were indeed very small differences, likely to be of minor consequence except in applications demanding 100% precise colour accuracy.

P.S. It would be interesting to see what happens to a ColorChecker Passport left out in full sun or in front of a strong UV source. (placing a series of thin strips of aluminum foil on the ColorChecker to cover half of each tile and then leaving it in sunny south-facing window could be a fun experiment!)

Considering the price of a new one, we might be waiting some time for a volunteer!

Seriously, though, it seems to be common wisdom that if the ColorChecker Passport is closed after use (to avoid unnecessary exposure to UV light) and stored under reasonable conditions, its practical life - for us amateurs and probably most professionals - will be considerably longer than two years...

[AstroDave]

P.S. It would be interesting to see what happens to a ColorChecker Passport left out in full sun or in front of a strong UV source. (placing a series of thin strips of aluminum foil on the ColorChecker to cover half of each tile and then leaving it in sunny south-facing window could be a fun experiment!)

Considering the price of a new one, we might be waiting some time for a volunteer!

I had the same thought about price!

If somebody has one of those 2 - 5 year old ones and figured it was due for replacement, (s)he could volunteer the old one for the experiment.

[swanlefitte]

Has anyone checked to see any of the following? Variablity with a card under different light sources after white set. A light source after being on a long time or years later. A lens aftet 5 years. Lens blog did a series checking lenses by the same brand and copy variation but not age.
These tests need answers before we can understand other tests. My needs are more for understanding light perception. If I photograph a squirrel under a tree the light should have a green tinge so accurate white is not natural. There a light triangle in the sense of accurate, natural, pleasing. To get one you compromise one or both of the others.

[BigMackCam]

If I photograph a squirrel under a tree the light should have a green tinge so accurate white is not natural. There a light triangle in the sense of accurate, natural, pleasing. To get one you compromise one or both of the others.

The thing is, even if you're using a 100% accurate neutral grey or white reference, the resulting custom white balance is going to produce neutral greys and whites rather than the green tint you're expecting in that situation. It just gives you a neutral starting point from which you'll need to adjust the colour temperature (and, possibly, the tint) manually in post-processing, in order to accurately reproduce the scene as you remember it, or according to personal taste...

[stevebrot]

The thing is, even if you're using a 100% accurate neutral grey or white reference, the resulting custom white balance is going to produce neutral greys and whites rather than the green tint you're expecting in that situation. It just gives you a neutral starting point from which you'll need to adjust the colour temperature (and, possibly, the tint) manually in post-processing, in order to accurately reproduce the scene as you remember it, or according to personal taste...

Yep...A neutral point is just the start. If you shoot the golden hour or the blue hour, keep in mind that it will likely not look golden or blue unless tweaked.


Steve

[swanlefitte]

I do like a neutral point. I often click between daylight, camera AWB and white card (when using one.) Artifical light is most likely when I want a good neutral. I wish the article would have covered the consistency across cards under various artificial lights. UV light being one problem mentioned above but uneven banding in the spectrum could be worse.

[photoptimist]

Has anyone checked to see any of the following? Variablity with a card under different light sources after white set. A light source after being on a long time or years later. A lens aftet 5 years. Lens blog did a series checking lenses by the same brand and copy variation but not age.
These tests need answers before we can understand other tests. My needs are more for understanding light perception. If I photograph a squirrel under a tree the light should have a green tinge so accurate white is not natural. There a light triangle in the sense of accurate, natural, pleasing. To get one you compromise one or both of the others.

Every light source has a color profile that changes with temperature, time, age, etc. It's a nasty problem.

Many off-the-shelf LED light bulbs have a color profile that varies over milliseconds. In many LED designs they use a blue LED with an over-coat on luminescent phosphors. The blue LED is energized directly by a simple circuit connected to the mains -- flashing at either 50/60 Hz or 100/120 Hz. The phosphors glow in the green/yellow/orange/red range to fill in the spectrum and the time while the blue LED is off. Using a high-speed camera, high shutter speed, or electronic shutter with this bulb design can have unexpected color results because the bulb is basically emitting a steam of blueish-and-yellowish oscillating colors!

Back in ye olden days, I worked on the color calibrations needed for a 10" x 20" film scanner for aerial photographs. The color profile of the light source -- a serpentine cold cathode fluorescent tube run on 20 kVAC -- varied significantly as the tube warmed up and the different phosphors changed their photon conversion efficiencies. We had an RGB sensor that monitored the bulb to do an overall, real-time calibration. But the tube also had a slight color profile difference from one side of the lightbox to the other due to temperature differences caused by how it was ventilated.

[BigMackCam]

I wish the article would have covered the consistency across cards under various artificial lights.

Does it really matter? I don't think it does. Like anything, you get to know your white balance and/or color profiling chart like any other piece of equipment. So long as you use the same one each time, it's all good. As I said previously, for most of us consistency is far more important than 100% accuracy.