Colour checker app?

Discussion in 'Digital Darkroom' started by rodeo_joe|1, Apr 16, 2021.

  1. Is there any liklihood that a self-luminous colour checker would work?
    Initially, I was thinking of a lightbox + dyed gel filter array affair, but a tablet with suitably calibrated screen then sprang to mind. Using the in-built brightness sensor to adjust to ambient light.

    Advantages - freedom from ambient lighting variations.

    Disadvantages - doesn't take account of ambient lighting variations. Only works in fairly dim light and fraught with all sorts of other potential issues..... but..... maybe?

    Non-starter or worth pursuing?
    Thoughts please.
    Last edited: Apr 16, 2021
  2. What you describe is contrary to the purpose of a color checker chart, which is to render true (i.e., daylight) colors under vaious ambient lighting conditions.

    Transparent color charts already exist in the form if IT8 slides for calibrating scanners. The effect is to standardize the light source, which is essential for a scanner but of trivial use in the situation you describe (unless your light box IS the light source).

    I have a small, folding X-Rite "Color Checker Passport" charts, separately for still and video standardization. I put the Passport on stage and short a short clip. I keep these clips on hand for venues I use regularly. It's especially helpful for matching diverse cameras, since each model seems to have unique color response. It is also essential for rendering Log Gamma profiles into something useful. Coincidentally, I nearly always use a fixed white balance, either approximate or measured. Most of my work involves musical performances, and brightly colored "diva" dresses mess up AWB royally.

    For standardization (aka calibration) purposes, the color chart must have an associated file which designates how it should be interpreted, patch by patch.
    digitaldog likes this.
  3. Non-starter.

    Have you ever read the original paper, "A Color-Rendition Chart," by C.S. McCay, etc al, published in 1976? I'm guessing that it can still be found online. It describes the color test patches and compares them, spectrally, with respect to the real world objects they were intended to simulate.

    Now, the color reproduction of modern color films, as well as digital cameras, is much better than the films of the 1970s, so you might say that testing is unwarranted. But say, for example, someone wants to "test" color reproduction of some object represented in the chart. The current chart allows you to compare, side-by-side, the actual chart vs the reproduced chart, under different light sources. You would not be able to readily do this with a backlit test chart.

    There's more, but I think this is a good starting point for consideration.
    digitaldog likes this.
  4. This sort of thing can be helpful even with digital, given the full range of different color temperatures
  5. Yes, I've got two of those Kodak colour-sep strips. Different editions with totally differently coloured patches. Plus a folding greyscale and RGBCMY card from Agfa-Gaevert (I guess that dates it!) and yet another Kodak colour swatch + greyscale folder.

    I was thinking something self-luminous might be useful as a portable D65/D50 reference, and a system colour checker. I.e. known colour inputs for calibration. Since, in theory, it could provide a wider gamut than is available from reflective pigments.

    I fully realise it wouldn't be much help for ambient colour balancing. Unless maybe by comparison with a reflective card.
    Ah, but the 'different light sources' is the nub of it. No use holding your reflective card up against your editing monitor if the ambient light isn't what your chosen colourspace has as a white-point.
    Whereas a colour-checker app would use known RGB values that could be directly compared to those of the reproduced colours. That being the point of an X-rite passport or whatever.

    Maybe OLED tablet display technology isn't at a stage where it's consistent enough for such a use? I don't know.
  6. A bit OT, but can one calibrate a tablet, say, iPad, (without a pc)
  7. You're kinda missing the point, which tells me you didn't read McCamy. Yet(?) Probably my fault for not explaining better.

    The original chart was made to mimic, spectrally (within a reasonable approximation), a variety of colored things. This allows you to test without actually having those "things." For example, from the 1976 paper, "The orange and yellow patches visually matched some samples of oranges and lemons, both freshly picked, not dyed. They matched when illuminated by Macbeth daylight with ultraviolet component, CIE Illuminant A, and a cool-white fluorescent lamp. Etc."

    So here's an example for the specific purpose of helping you understand my point. Say you wanna see if your film/paper system, or whatever, can make a decent printed reproduction of a similar orange and lemon. So you photograph the chart (or you photograph the actual orange and lemon if you prefer). Now you make a print. Now, as I said, you can compare under different light sources. You can hold the print, the Macbeth ColorChecker, and perhaps even the original fruit, in a color-viewing booth and compare them. Or you can compare them under whatever fluorescent lamp you like, even the (horrible-for-color) eco-friendly style, like most CFLs. Or you can carry them outside and compare. That's what I mean by comparing the original chart vs reproduction under different light sources. You wanna see how your reproduction holds up under different conditions.

    Now, I can guess what you're thinking, "Why bother with the ColorChecker? Well, by using a standard reference chart, the ColorChecker, you can do this testing when the fruit is out-of-season. Or if you have a remote customer you don't have to mail the fruit to them.

    I could say a lot more but I don't think today's, in general, has too much interest.
    digitaldog likes this.
  8. digitaldog

    digitaldog Andrew Rodney

    The ColorChecker Pages (Page 1 of 3)

    The ColorChecker Color Rendition Chart, this very well known chart with an array of 4 x 6 color patches, is an icon of the imaging industry. It was formally presented in a 1976 article by C. S. McCamy and his colleagues from the Macbeth Company, a Division of Kollmorgen Corporation at the time:•C.S. McCamy, H. Marcus, J.G. Davidson, “A Color-Rendition Chart,” J. Appl. Phot. Eng., Vol. 2, No. 3, Summer 1976, pp. 95-99, Society of Photographic Scientists and Engineers (Now called “The Society for Imaging Science and Technology (IS&T)”; of the article are not available in IS&T online store but they can be ordered by sending an e-mail to Customer Service.A pdf of the scanned article is freely available from the Rochester Institute of Technology Web site (Copy/Paste this link:
  9. X-Rite has an app for iOS devices which can be used with one of their Color Munki devices to calibrate an iPad. If your iPad uses a Lightning port, you need a Lightning to USB A adapter. Newer iPads have a USB-C port, which also needs a USB-A adapter. A profile is created which is used by some imaging software, including Lightroom Mobile.
  10. All a bit out of date now.
    That was fine when I was darkroom printing colour negs, but the last time I did that was nearly 20 years ago.

    Almost nobody compares prints to a bit of coloured card any more. No matter how expensively made that bit of coloured card is.

    I used to shoot a frame of a colour card on each film, but that was near useless if the lighting conditions varied for the rest of the frames - which they frequently did. And a tiny colour checker stuck in the corner of each frame was too small to be of use in those days. There's no eyedropper tool on an enlarger!

    Anyway. All of this totally gets away from whether there's any point in having an ambient-independant set of portable colour patches.

    I can see a use for a one-time end-to-end system calibration. Maybe not a lot of use beyond that, and with different input/output brightness ranges to contend with or compensate for.
  11. Too late to edit, but to add the above.

    The big advantage I can see of a luminous target is its ability to present a greater brightness range than a reflective target. Maybe 300:1 against the 125:1 (if you're lucky) of a semi-matt surface. While a gloss surface presents all sorts of specular reflectance issues. Unless you use concave-contoured patches, which are expensive to produce.
  12. Most color charts have a glossy black patch, and possibly a white patch too. They represent the blackest black (and brightest white) you can easily obtain from a reflective chart. A glossy patch is actually less affected by ambient light than a matte one, if you use care to avoid specular reflections of light sources. Reflections are best controlled by lighting angles, not polarizers, even when copying artwork.

    I calibrate my display and printers, but rarely a camera for stills. Video is different, simpler in some respects but less well defined. Rather than RGB you adjust R, B, and luminosity, and the process is highly subjective. The test charts are simpler too, with 3-5 grey patches and 5-24 color patches. The most objective analysis relies on vector and cascade displays, but that's never enough. In the end your eyes are the best tool for color and exposure balance, especially when matching cameras. Not many people have a REC709 monitor, so what your customers actually see is a wild card. For broadcast, you have a legal obligation to keep luminosity under 100%, and blacks above the interframe basement. It's a lesson in humility.
  13. I've made an absurdly large dynamic range grayscale target by back-illuminating various step tablets, some of which had ND filters behind them. Worked great for lab testing. I can imagine the illuminated target described would also be good for some kind of lab testing, but not so much for practical picture-taking applications. AFAIK, those Kodak color control patches are for graphic arts and not as useful as something like a Macbeth color checker, costly though it is, because it's been well characterized.
  14. Not that it matters much, but it should also be recognized that some light sources do not have a continuous spectrum, but have gaps (just as astronomers use...).

    incomplete spectrum
    CD 21 Spectrum of a fluorescent e.jpg

    If not all the colors are present, it can be extremely hard to "correct" since it's hard to brong back something that is no longer there at all.
  15. Ah, well. That's the 'joy' of a tri-colour system. In theory you only have to juggle with 3 balls.

    However. Here are a few full sunlight spectra as shot by an assortment of digital cameras:
    Top is the spectrum as viewed by eye through a simple slit spectrometer, and below that, what the digital cameras capture.

    Whoops! Where have all the chromatic yellows and cyans gone?
    Gone to graveyards every one.

    I wish colour accuracy wasn't always being sacrificed on the altar of saturation.
  16. So I'm curious as to how you get the image, as viewed by eye? Photographed on film? Or a "higher grade" digital? Or is it synthetic?

    I oughta try that test myself.

    It'd also be interesting to see the results from a Foveon sensor camera; I'd probably be willing to bet that they don't have any trouble with spectral yellow.
    Last edited: Apr 20, 2021
  17. digitaldog

    digitaldog Andrew Rodney

    This is what most digital cameras actually capture:

  18. digitaldog

    digitaldog Andrew Rodney

    "All generalizations are false, including this one." -Mark Twain
    Accuracy (undefined word used; colorimetric accuracy?) isn't always sacrificed on the altar of saturation. Its why perceptual rendering intents can be rather interesting to study colorimetrically IF you have the tools. You can even lump in actual color accuracy into the discussion. ;)
  19. A simulation I'm afraid.
    Nothing I have can actually capture those spectral yellows and cyans.

    I gave up using colour film years ago, but I suspect it might do a better job - in some respects.

Share This Page