German camouflage colors.

[Patrick]

Here is my derivation of the Fokker D.VIII wing stain colors.

[Patrick]

For comparison, I used Jan's colors for the night pattern and produced it as both an untextured and a textured swatch. My original derivation and the new one using Jan's colors are both posted here (first two attachments are mine, second two are Jan's).

Dan, are either of these anything like what you intended for this pattern? How do they vary from your intended colors?

[Dan_San_Abbott]

Patrick:
Yours pattern is very close to correct pattern. I don't know what happen with Jan.
Blue skies,
Dan-San

[Patrick]

I don't know what happen with Jan

I have a rough idea, because I have the same problem when attempting to match color from a chip chart; I tend to pick them too light when you mix them into the actual pattern.

I suspect that the "A" Blue in the night pattern is too bright; at least it looks so to me. If you have other Methuen colors you'd like to see for this pattern (or any other) just post & I'll set them up here.

[JohnFitz]

Patrick,Dan-San, In the #51 post example of the D.VIII wing stain would you say these colors should be in the four color fabric also? I know the light passing through the fabric would make them look lighter however. Cheers!

JohnFitz

[jumpinjan]

I spent more time comparing the night pattern and made some slight changes (mostly darkened some). The link is
http://www.sunbeamalpine.org/jans_pi...All_colors.jpg

[Langdon]

Gentlemen,

Congradulations on a very interesting thread, it seems that colour comparison, particularly electronically, is a difficult subject. I have noticed that if I use the eyedropper tool in photoshop to check your colours I get slightly different values to yours, what would be the reason for this (probably a silly question but I know little about this subject)? The RGB figures I get from the top row of Jan's latest colour card are 97, 104, 110. 114, 95, 127. 53, 60, 70. 108, 114, 76. 142, 100, 60.

Langdon

[Patrick]

Patrick,Dan-San, In the #51 post example of the D.VIII wing stain would you say these colors should be in the four color fabric also?

My answer is No - the #51 post shows stain samples that are applied to plywood & varnished over (although some on the Fokker D.VIII Wing thread speculate that the color could actually be part of the varnish -- added as a "tint"; I have no opinion on this).

Four-color fabric colors are different, although the stain and fabric colors may have been intentionally picked to be harmonious - Dan would know more about this than I. Look further back in the thread for four-color fabric reference cards for comparison.

[jumpinjan]

I noticed that too. Mine are:
97/104/109, 114/94/126, 54/60/70, 107/114/76, 141/100/59
What I did was to use the Corel Draw's color rollup to make the illustration, then export it as a .psd (Photoshop) file and then use Photoshop to size it and convert to .jpg. I have been using that combination of applications for about 10 years now, but now there might be better ways of creating the jpgs.

note - My numbers are on the 6meg psd file, but I do get EXACTLY the same numbers as you Langdon with my posted jpg.

[Patrick]

I have noticed that if I use the eyedropper tool in photoshop to check your colours I get slightly different values to yours, what would be the reason for this

Because if you're using Adobe PhotoShop to pick the colors off of a reference card, your internal color space profile is probably Adobe98. My reference cards were developed in PS, but I use a profile called BruceRGB. Jan uses Corel 8 and I don't know what the default color space profile is -- possibly sRGB. There are literally a dozen internal color space profiles pre-defined and PS gives you the capability of defining your own. You use one vs. the other depending on what you're targeting for the color.

What is not clearly understood is that RGB is a device-dependent color code. That means that the exact RGB numbers you get when you use the eyedropper depend both on the internal color space profile AND on your monitor AND on how your software handles the mapping from internal color space to monitor space. That's why we can't just provide RGB numbers and expect people to get good replication on their equipment -- what two people see on two different systems for the identical RGB numbers can be very very different.

That's why I provide the Lab numbers on my reference cards (and Jan, you should start doing this too). Lab is a device-independent color model & is widely used to communicate color data in a precise manner. Lab was developed in 1931 as a mathematical model of how humans perceive exact hues of color under tightly specified lighting and viewing conditions. Proof of how good a job was done in defining this model is that it is still widely used and recognized today as an international standard for color. Despite years of attempts, no one has ever come up with a system that is superior. It has no dependencies on any physical device, rather, physical devices have to support the Lab model. That's why Allan's color samples on the Fokker D.VIII Wing thread look so close to mine -- he created them using the Lab numbers, not the RGB numbers.

Take the Lab numbers from one of my reference cards and plug them into the Color Picker. Note the color that you get -- it should be very close to the chip on the card. Note also that your RGB values (in the Picker) are probably quite different then mine. This simply reflects the color space differences between your equipment and mine. Neither is "more correct" than the other -- your's is correct for your system and mine is correct for mine.

The only reason I provide the RGB numbers is as an approximation for those people who don't have software that supports Lab. The RGB numbers I provide for in-gamut RGB values come from the GretagMacbeth Munsell Conversion tool and represent a linear conversion from Lab to sRGB color space. If the Conversion toll tells me that the Lab values generate an out-of-gamut RGB color, I provide the RGB values from my system (taken from the Color Picker in PS). This is a really bad set since it's a device-dependent set for an approximation of a color that can be printed but not displayed on an RGB CRT.

Got that? Clear as mud, right?