Regarding Exact Color Syndrome, the late Al Savage, VP of maintenance at DC Transit often talked about matching paint for what was the Capital Transit one-man paint scheme used on the PCCs and standard cars. He admitted over the years that the "electric blue" became more green, and the cream "V" on the front tended to go more tan. Likewise the striping at the belt rail varied from dark red to brown. It was based on the current paint mix and what they had in the shop. At NCTM we finally settled on the fleet color "Norway Green" for the "electric blue".

Al liked to tell the story of the color for the PCC trucks. Originally painted a medium gray, the trucks were hard to keep clean. Al had the paint shop mix up a new color he called "dirt" to match the dirty shade of gray and the problem was solved.

We maintain some drift cards and other paint samples and codes for the car collection. In some cases we have simplified things, using a single near match for two different shades of the same color. Thus "Hudson seal" serves as the truck color "dirt" as well as a very close match for the truck frames and hardware on several other cars. The same is true for "tuxedo cream" and "Dixie gray" which appear on several cars that have been repainted over the years.

Our approach is not an exact science and seems to be in keeping with what Al Savage described as a Best Practice from the transit company.

Wesley

I was thinking about this very issue today.

In an effort to help future historians, I carefully documented a string of four Union Pacific Locomotives. You are welcome to use them on future UP restoration projects to establish the exact shade and tone of UP Armour Yellow and Cascade Grey. Just doing my part for railroad preservation, no need to thank me..

I have the ability to shoot in raw, calibrate my images using a color target in the field, which is then adjusted using special software. Using a calibrated monitor and printer profiles, I can then print quite accurate photos.

But I didn't have time for all of that today, so I shot this in jpg with a point an shoot, and I'm uploading them to the web, so your browser can be sure to screw up the profile.

6963!?! A 6900 Isn't that one of those DD40's???

(For the newbies, no, it's not... A Google search indicates it's a AC6044CWM, but I'm not enough of a diesel geek to know for sure...)

Also note the subtle variations in colors of the two american flags.

Not to mention the slight color shift in 5581's air conditioner.

-Hudson

On a more serious note, observe how all of the paint schemes on this random string of locomotives vary. Two have flags, two don't. The "Building America" slogan and shield are in different locations.

OK, but what about the top two, aren't they the same? No, they differ as well. On one, the top red stripe has a zigzag and it drops down. The other one does not.

Now, ask yourself this: How hard is it to imagine some future version of RYPN and an argument over "No, the shield was in front of the flag!" "No, it went further back on the hood!"

This is on the UP, where a casual observer could easily conclude everything was painted in the same manner except for the flag units. Now imagine BNSF with all of its variation and patch jobs!

DDA40X numbering ended at 6946.

There are and will be many variations on paint, location of this or that and not to mention the color shades themselves....fading, road grime and patches and repairs. But I am certain that the debate will rage on.

Please keep in mind that new organic paint pigments may fade differently than older pigments. A lot of improvements in yellow pigments have been made in recent years. So todays Armour yellow paint may fade differently than paint applied five years ago. What you usually see on railroad equipment is faded paint.

In the old days all yellow paint and many green paints had chrome yellow pigment. Chemically it is lead chromate so it became a no-no for two reasons. Old paint formulas are worthless because many of the pigments are no longer available.

Now the best you can do is color match an existing locomotive and then try to keep
your new paint out of the sun as much as possible. There are NO exact colors today.

So, the above, all accurate, defines the problem but not the solution... (the UP photos start to document the problem... which in itself is a solution)

As preservationists we need to document color and paint and lettering... I would suggest weathering as well... (aka the UP photos) At each step... as well as possible...

The fact that there are lots of variations, and that paint fades makes the problem more complicated, but doesn't excuse sloppy documentation...

Take photos... sample paint layers... document what is found via photos and matches (PMS is more portable... Munsel is a better system, neither hits the dark reds and greens we deal with with older equipment) Write up findings... If you can take paint samples and protect them. Make your findings available.

The more information we preserve (save and organize so those following can find it) the better... and keep the colors documented as exact as possible... with notes about how you found them... the restoration crew can then make good decisions about what color to use... for that day in Oct 2009....

Randy

I have a friend who is a very talented HO scale diesel modeler. He superdetails and paints his models based upon a locomotive's appearance on a given day or week. Seems like that is what is being advocated here in 1:1 scale.

Rob

Addendum to my previous post: Do NOT try to color match to photos. The color rendering system in photos is very different from paint in daylight. Photos are very useful for determining that stripes were painted red and the pattern of the stripes. Don't try to determine the shade of red paint used.

Human vision is not capable of making absolute color judgement similar to perfect pitch for musicians. Human vision is capable of making very accurate color comparisons, as many as 10,000 according to some authorities.

However human vision integrates colors received by the eye. Literally dozens of different color mixes can appear to be an identical color when viewed.

"Do not try to color match to photos".

Without getting all photography geek on you, I will second that. First of all, there's a thing called "white balance" that can greatly affect how the image looks. Then you have the variance the monitor and the printer as well. Serious photographers spend lots of time calibrating and using color targets to reduce the problem, but it's still there.

Also, as mentioned, photos render colors differently that pigment. Then there's the whole "index of refraction" problem that means the same color looks different on different objects. The deeper you did, the more complicated it gets.

Here's a mind blowing page full of color and perception illusions.

You will find some of them very difficult to believe. For a few of them, I've gone so far as grabbing screen shots and opening them in photoshop to verify what they're saying. As an example, it's hard to believe a color that looks brown in one part of an image and yellow in another is the same exact color.

http://www.lottolab.org/articles/illusionsoflight.asp

That reference is the most spectacular I have ever seen. Let anyone who says "photos don't lie" look at it.

If you would like a good demonstration of shortcomings of color photos, make several photos of the same scene using different exposures. Do not change the f stop. Include a number of different colored objects in the scene. Then print or view the photos so that they appear to have similar brightness. Notice what has happened to the diferent colors in the scene. This is what happens to an object in a color photo that is partially in direct sunlight and partially in the shade.

Incidentally the two main pigments in Pullman green or coach green were chrome yellow and carbon black. So coach green actually is a very dark yellow.

Munsell or functional equivalents are the only ones that work long-term. That is because they rely on the physics of light in absolute terms.

All the other systems - including Pantone - rely on specific, made pigments continuing to be available and continuing to be the same color. (how do the pigment makers get an absolute reference? Munsell!)

And pigments go out of production - ask your paint man, he'll tell you. When a pigment system gets changed, all those paint-store chips with colors like "Honeysuckle Rose" go in the trash.

You can't substitute an ancient pigment with a blend of modern pigments. Virtually all formulas are a blend, and your substitute pigments will interact in weird ways. What can I say but to try it? Go into your shop and find one of your custom "yellow"s that is actually a blend of several pigments, and a "red" which is likewise. Mix a teaspoon of each, do you get the robust orange you expected? No, you get a surprisingly dull and washed out brown.

As far as photos, forget about it. Here. Here's a photo of the car next to some Munsell chips, that's as close as we get to "not wrong", and can YOU tell? Photoshop is little help