Improvised color chart - How reasonable is this?

[Romanko]

Imagining that one can download a color chart of any serious value and print it themselves is a pipe dream.

Reflective targets from Coloraid and several other manufacturers are printed on photographic paper and measured either individually or for each batch. I don't know what methods they use to produce them, probably a device similar to a film recorder. These targets are considered adequate for calibrating a consumer scanner or a digital camera.

As I mentioned it all depends on the application and your requirements. While printed targets and maybe even DIY ones are sufficient for obtaining a simple matrix calibration profile for your digital camera (provided it even requires one), professional reproduction of art work calls for a very careful approach and proper equipment, not to mention years' of experience, taste and artistic vision.

 

[koraks]

I don't know what methods they use to produce them, probably a device similar to a film recorder.

For the reflective targets, it wouldn't surprise me if they turn out to be regular minilab prints, or even farmed-out lab prints. It's just digitally exposed RA4 paper; nothing special. Although of course that is special in a way, because the dyes in these papers are actually pretty high-quality for this purpose. This becomes apparent if you look at their spectral density, which is very nice and gradual with no ugly lumps & bumps & peaks.

For the transmissive targets some kind of film recorder device is evidently used.

 

[Bill Burk]

We’re talking about walking into Ace and picking a bunch of free paint chips, gluing them onto a card and having a test card.

Sounds fun.

 

[Romanko]

We’re talking about walking into Ace and picking a bunch of free paint chips, gluing them onto a card and having a test card.

Paint manufacturers specify color as RGB values. Anyone knows what color space they use?
Any Argyll CMS users here? How hard is it to specify a custom color checker file in Argyll CMS?

 

[DREW WILEY]

Of course one can jerry-rig a makeshift color target using paint chips. But it's never going to be balanced and accurate. Each misstep along the workflow process tends to compound itself : lack of neutrality or pure chroma in the original target, combined with reproduction idiosyncrasies inherent to EVERY kind of color film and print medium, or scanning and screen variables. The better your starting point, the higher the odds of being closer to the bullseye at the endpoint.

I find this very important whenever calibrating the starting colorhead settings of a new batch of RA4 paper, for example. Specific images can be manipulated creatively as needed; but if you don't have a defined objective starting point, a lot of paper might get wasted getting the result you hope for. Likewise when attempting to do any truly objective color film test.

The classic MacBeth Color Checker Chart is now marketed by Datacolor, which has a whole suite of tools allowing you to integrate your total workflow.

A photo lab can no more mass produce accurate color targets on RA4 paper than a paint store can make an accurate 18% gray paint. Otherwise, I just don't get it. People will spend thousands of dollars on some new lens or camera they don't even really need, or on the latest Smartphone, but aren't willing to invest $50 in a reliable color target? Those things save you money in the long run.

Romanko - I partnered with paint and pigment manufacturers for a long time - the very best of them - and believe me, the suggestion that real-world paint pigments can be reduced to simple RGB parameters is hopelessly simplistic. Yes, these manufacturers use basic four-axis CIE models to program automated tinting machines. But the fact that a dozen or so pigments are typically involved informs one that they are never dealing with a true set of CMYK process colors. Inkjet printing is analogous. And a margin of variability is always built in. Then you've got the significant issue of metamerism.

Paint chips aren't actual paint. And the fact they're handed out for free means they're relatively cheaply made, and vary quite a bit batch to batch. I've got in storage somewhere a personal set of architectural-grade color samples, each about a square foot in size, filling five thick volumes. These were expensive to make and never available to the public - only color pros. But even these wouldn't necessarily precisely represent how the final paint itself would turn out. There are just too many variables.

 

[Romanko]

Of course one can jerry-rig a makeshift color target using paint chips. But it's never going to be balanced and accurate.

I believe we all agree on this. We are comparing a DIY color chart of questionable quality and inaccurate values to no calibration at all.

 

[Wolfram Malukker]

I've given up on this line of questioning, too many here are locked in on "it can never work" and aren't interested in any other options. The ones who have contributed useful information have made good input, and I will look into other options.

 

[Romanko]

I've given up on this line of questioning, too many here are locked in on "it can never work" and aren't interested in any other options.

Please understand that the skeptics have professional training in color science and decades of experience (Thank you @DREW WILEY for contributing to this discussion.) Their standards and definitions of "work" are much higher and stringent than those of an average hobbyist like myself.

If you are looking for a simple approximate option, there are a few to choose from:
https://www.youtube.com/results?search_query=diy+colorchecker
They "work" to a degree. Whether it is sufficient or not is up to you to decide.
A proper color checker is the price of a several rolls of color film.

 

[DREW WILEY]

How many hundreds of examples have you seen of a box of Crayola crayons being used as a color target, especially in color film ads? Well, all those colors somehow come out looking bright and snappy in those ads. But then when you use that film yourself, real-world subjects turn out differently.

Wolfram - I'm not locked in. Rather, I've "been there, done that", and then some, and know exactly what to expect. I've tried lots of alternatives myself. I even know what it takes to actually batch up truly neutral 18% gray paint; I've done it, and it isn't simple. If someone wants to run all over some large city collecting various brands of paint chips, hoping these will cumulatively approximate a set of properly saturated colors or neutral grays, they'll probably spend as much in gasoline as what a professional color chart costs. I not only trained color matching personnel, but had a sample library of color chips from all kinds of manufacturers - both domestic and European on hand, industrial, Govt spec, residential. Sure, the paint chip route is certainly better than nothing, and might constitute a fun project, but is otherwise more of a detour than a real solution.

 

[Bill Burk]

Of course one can jerry-rig a makeshift color target using paint chips. But it's never going to be balanced and accurate. Each misstep along the workflow process tends to compound itself : lack of neutrality or pure chroma in the original target, combined with reproduction idiosyncrasies inherent to EVERY kind of color film and print medium, or scanning and screen variables. The better your starting point, the higher the odds of being closer to the bullseye at the endpoint.

I find this very important whenever calibrating the starting colorhead settings of a new batch of RA4 paper, for example. Specific images can be manipulated creatively as needed; but if you don't have a defined objective starting point, a lot of paper might get wasted getting the result you hope for. Likewise when attempting to do any truly objective color film test.

The classic MacBeth Color Checker Chart is now marketed by Datacolor, which has a whole suite of tools allowing you to integrate your total workflow.

A photo lab can no more mass produce accurate color targets on RA4 paper than a paint store can make an accurate 18% gray paint. Otherwise, I just don't get it. People will spend thousands of dollars on some new lens or camera they don't even really need, or on the latest Smartphone, but aren't willing to invest $50 in a reliable color target? Those things save you money in the long run.

Romanko - I partnered with paint and pigment manufacturers for a long time - the very best of them - and believe me, the suggestion that real-world paint pigments can be reduced to simple RGB parameters is hopelessly simplistic. Yes, these manufacturers use basic four-axis CIE models to program automated tinting machines. But the fact that a dozen or so pigments are typically involved informs one that they are never dealing with a true set of CMYK process colors. Inkjet printing is analogous. And a margin of variability is always built in. Then you've got the significant issue of metamerism.

Paint chips aren't actual paint. And the fact they're handed out for free means they're relatively cheaply made, and vary quite a bit batch to batch. I've got in storage somewhere a personal set of architectural-grade color samples, each about a square foot in size, filling five thick volumes. These were expensive to make and never available to the public - only color pros. But even these wouldn't necessarily precisely represent how the final paint itself would turn out. There are just too many variables.

The secret is to add orange.

 

[DREW WILEY]

Orange juice in lieu of an orange negative mask? The actual secret is understanding how paint manufacturers do indeed intend their sample chips to reasonably approximate their own actual formula colors; but those are restricted to certain parameters, including pigment affordability and covering opacity. Pure primaries and tertiaries are not the goal. Most brands here in the US work with UTC's ("universal tinting colorants") largely dependent on relatively cheap oxides and non-neutral blacks. Don't expect chip colors to represent expensive quinacridone reds and magentas, for example, except in the most expensive paint brands themselves. Then you've got the issue of cyclic trends, where certain decades might favor muted colors in general, rather than bright crisp ones; so nearly all the given chips will intentionally look dirty. Working with Pantone color sets is even more dicey, though it does provide more of a common denominator system.

All this is a helluva lot of fun to juggle around if you're being paid to do that as a pro color consultant; but from a photographic standpoint, it really isn't all that much better than the good ole box of Crayolas method.

 

[koraks]

too many here are locked in on "it can never work" and aren't interested in any other options.

I admit I'm very skeptical, but personally I'd be very interested in being proved wrong by a working color management approach (even if it's seat-of-the-pants quality). I like the simplicity of your paint patch idea and encourage you to try it - but I'm also saying this on the basis of many, many experiments trying to match colors and densities by eye in a variety of applications. So I'm well aware of the pitfalls from an entirely practical, amateur perspective, much like yours. Hence my skepticism!

At the same time, I've also done the dance with a simple monitor calibration tool and ICC profiles for my printer, and on that basis I know that with half a days work and a $200 investment you can do exactly what you're trying to do in a no-fuss, dependable way.

I would definitely recommend starting with the paint chips first, and then if (and only if) you feel the need (e.g. because you purchase a new monitor and thereby accidentally 'break' your 'calibration' chain), re-consider the suggestion of profiling in the regular fashion.

 

[MattKing]

In the meantime, you can always use your paint chip colour chart as a reference, to be included at the edge of a lot of scenes as a starting point for matching.
Not as good as a Macbeth product, but better than a lot of other things.

 

[_T_]

I don't use any color charts, I don't calibrate my monitor, and I haven't profiled my scanner, yet whenever I send my files to the lab for printing they come back looking the way I intended.

On the other hand I used to work in a studio with a lab that empirically managed their color at every step and so I know how even a small problem at any step can cause major discrepancies in the final prints.

If you're going to make improvements to your color accuracy, beyond doing nothing, by making empirical measurements, the only way to gain advantage over doing nothing is to take all of the steps, and if you're going to do that it would be to your benefit to do it correctly. Otherwise you end up in a situation where you're spending time continually chasing down problem after problem that you simply wouldn't have if you just didn't do anything in the first place.