Spitfire cockpit green

[mhaselden]

A doctor used a refractometer on me once...made my eyes water (and the device had nothing to do with my ocular capabilities!).

[Max Headroom]

Are we in terminal thread drift?

Trevor Duck Egg Green

[expositor]

That just shows the working and provides the base data rather than just making an assertion. The first three values are from the L*a*b* colour space which covers all perceivable colours and is closest to the way humans perceive colour. Most measuring devices measure colour in L*a*b* and the Federal Standard publishes those values for each of their colours (as does BSi). In essence they are the "instructions" that allow someone to "build" the same colour without having to see it. For showing the colours on screen devices like computers the values must be converted to RGB - Red, Green, Blue - which inevitably causes some minor but usually negligible distortion (dependent on the calculation methodology used). CMYK is generally used for printing colours consistently but most colour management, cgi or photoshop type programmes allow for RGB and/or CMYK values to be input to produce colour which I thought might be useful so that people can "see" the colour for themselves or to reproduce it on their own systems. Those who might find it useful are welcome to it and those who do not need not worry about it.

The difference quantification of 0.92 was calculated using the DE2000 formula. This formula is the most recently recommended by the Commission International de l’Éclairage (CIE). The lower the number the closer the match - a difference calculation of 2.0 or less usually indicates a close match, but the figure is drawn from cumulative values so the direction of colour shift must be quantified by description and/or observation. Generally any calculation beyond 5 makes a comparison match of questionable value but the variance of applied wartime paints is often up to about 5.

I showed the working to reassure John that the FS colour has not changed because the values officially assigned to it are consistent and when converted into other media produce the exactly similar colour - therefore it is peer verifiable.

HTH

Nic

So there is actual science in the FS color classification system? If that's so, why do we see colors clearly more blue in the 6's, greys in the 5's, greens in the 0's, browns in the 4's, etc? This is what I alluded to many, many moons ago when I posted that there seemed to be more art than science in the FS system, and not as you misinterpreted that I believed you didn't use 'science' to pick FS color matches for specific colors in other numbering systems. It would have been more 'sporting' to see a response other than "no comment" in this venue, rather than a derisive one elsewhere on the web, but we can still be friendly, no?

Anyway, I hope you don't think you're wasting your time if you decide to enlighten me. Not to be patronizing, but I, as we all do, appreciate you sharing your expertise, and I always look for your posts when I lurk here...like everyday....

Cheerio!

Edited August 12, 2014 by expositor

[Nick Millman]

The Federal Standard presents the colours used by US Government Activities for the purpose of colour selection, colour matching and quality control. It is not a colour classification system like Munsell or Methuen but a limited catalogue of specifically selected colours grouped together in "sets" by predominant colour grouping. The "sets" are identified by the second digit of each colour's five-digit code as follows:-

0 Brown

1 Red

2 Orange

3 Yellow

4 Green

5 Blue

6 Grey

7 Misc

8 Fluorescent

The first digit identifies whether the colour is gloss 1, semi-gloss 2 or matt 3. So for example FS 16350 is a gloss grey. The last three digits identify the individual colour and also represent its approximate reflectivity on a scale of 000-999.

The science part comes in because each colour also has a specific value in L*a*b* and a specified set of pigments used to create the chip. So FS 16350 has values as follows:-

L*a*b* = 58.36 0.31 11.11

CMYK = 40 40 60 0

sRGB = 147 140 121

Which allows that specific colour to be recreated in those formats and also functions as a control mechanism by which matches can be verified.

The pigments used to create the FS 16350 chip are Rutile Titanium Dioxide (white), Yellow Iron Oxide, Carbazole Violet and Carbon Black (Blue Shade).

That's about it.

Nick

[gingerbob]

[Oops, Nick snuck in.] Huh? Sounds like "damning with faint praise" to me! One thing I'll say for Nick (I could say more for him, but another time perhaps ) is that whether you can relate to the science or not, he generally does a good job of putting his point into "Plane English" (deliberate misspelling) as well. I'd rather have to engage my mind a bit than just get over-simplified answers.

Nick: Unfortunately I am both pedantic about colour science and relaxed about modelling colours which must seem a confusing contradiction at times.

I get it. I like to understand as well as possible (actually, I'm probably fooling myself), then I can choose my own personal "works for me" level.

bob

Edited August 12, 2014 by gingerbob

[expositor]

The Federal Standard presents the colours used by US Government Activities for the purpose of colour selection, colour matching and quality control. It is not a colour classification system like Munsell or Methuen but a limited catalogue of specifically selected colours grouped together in "sets" by predominant colour grouping. The "sets" are identified by the second digit of each colour's five-digit code as follows:-

0 Brown

1 Red

2 Orange

3 Yellow

4 Green

5 Blue

6 Grey

7 Misc

8 Fluorescent

The first digit identifies whether the colour is gloss 1, semi-gloss 2 or matt 3. So for example FS 16350 is a gloss grey. The last three digits identify the individual colour and also represent its approximate reflectivity on a scale of 000-999.

The science part comes in because each colour also has a specific value in L*a*b* and a specified set of pigments used to create the chip. So FS 16350 has values as follows:-

L*a*b* = 58.36 0.31 11.11

CMYK = 40 40 60 0

sRGB = 147 140 121

Which allows that specific colour to be recreated in those formats and also functions as a control mechanism by which matches can be verified.

The pigments used to create the FS 16350 chip are Rutile Titanium Dioxide (white), Yellow Iron Oxide, Carbazole Violet and Carbon Black (Blue Shade).

That's about it.

Nick

Thanks Nick for taking the time. So I stand corrected on the science thing, but what about pre-computer? Do those values you listed have a direct correlation to an FS number, sans computer reproduction? You also state that the 000 to 999 numbers relate to 'reflectivity.' For me the amateur, reflectivity would refer to the 1, 2, 3 for gloss, semi, and flat, and, or the degree to which a color would reflect solar heat independent of its actual hue. The aforementioned 000 to 999 numbers, so I believed, referred to each color's darkness to lightness, the lower the three digits, the darker the color, so reflectivity in that vein is a bit confusing to the ignorant among us...or is it just me? I have believed that the actual three digits following the color number is an arbitrary one so they can add new colors, (or am I wrong about that too?) which is why I ask again, why are some colors predominantly(?) blue to the eye (well, mine) and given an FS number in the 6's; some looking predominantly grey, given one in the 5's, an obvious(?) cream color also in the 6's, etc?

Thanks again!

Edited August 13, 2014 by expositor

[Nick Millman]

Thanks Nick for taking the time. So I stand corrected on the science thing, but what about pre-computer? Do those values you listed have a direct correlation to an FS number, sans computer reproduction? You also state that the 000 to 999 numbers relate to 'reflectivity.' For me the amateur, reflectivity would refer to the 1, 2, 3 for gloss, semi, and flat, and, or the degree to which a color would reflect solar heat independent of its actual hue. The aforementioned 000 to 999 numbers, so I believed, referred to each color's darkness to lightness, the lower the three digits, the darker the color, so reflectivity in that vein is a bit confusing to the ignorant among us...or is it just me? I have believed that the actual three digits following the color number is an arbitrary one so they can add new colors, (or am I wrong about that too?) which is why I ask again, why are some colors predominantly(?) blue to the eye (well, mine) and given an FS number in the 6's; some looking predominantly grey, given one in the 5's, an obvious(?) cream color also in the 6's, etc?

Thanks again!

"what about pre-computer? Do those values you listed have a direct correlation to an FS number, sans computer reproduction?"

Yes, L*a*b* derives from a 1931 system for measuring and identifying colour. It is both an input and output for real colour, e.g. the colour of a paint surface. So for example if you mix up a completely arbitrary paint colour using hobby paints the resulting colour can be measured and expressed in L*a*b* values. RGB systems go back to the 1920s but are based on colour theories developed in the mid-19th Century in the earliest colour photography experiments. I am less familiar with CMYK which is used in colour printing but I believe it pre-dates computer technology.

"...so reflectivity in that vein is a bit confusing..."

Without going into the differences between specular and diffuse reflection, colour reflectivity is inherently related to the lightness and darkness but is usually expressed as a percentage. For example pure white is 100% and pure black 0%. The FS system is untypical and does not correlate exactly to that. Since each colour has a unique identifying number there would be an issue for colours having identical reflectivity. The FS specification document itself says this:-

"S3.2.3 Last Three Digits. The last three digits of the color number are assigned in the approximate order of increasing reflectance."

An approximation can be gained by dividing the last three digits by 100, thus 16350 = 35%, but it is not exact and I have found that non-FS colours that match FS colours very closely do not share the reflectivity expressed by the last three digits.

As to the colour sets I don't know the basis for how the colours were assigned by FS but if it was visually then issues of subjective perception would apply, especially for colours occupying subtle colour space - green-browns, warm greys, pale blue-greens, etc. The FS specification documents suggests no particular precision in this (my emphasis):-

"S3.2.2 Second Digit. The second digit of the color number indicates an arbitrarily selected color classification grouping."

Several of the FS greys (6) are indeed Munsell Blues but 16160 and 16165 are Munsell Yellows! Some might legitimately say that 16165 looks "greenish". Whether that assignment represents an intended usage I couldn't say, but a pattern of co-ordination can be seen in the fan deck layout light to dark. Even so each colour has a "fixed" and published L*a*b* value that identifies it beyond doubt.

The problem with using FS as a matching system is that many colours in other systems and paint colour standards do not have precise equivalents. So a visual near match ought to be quantified but seldom is. Therefore the nearest colour in the FS system, often arrived at by subjective visual comparison, ends up becoming the colour for modelling purposes and one sees anachronisms like "The wings (of a 1930 aeroplane) were painted in FS 13538". The system is limited but very visual which makes it easy to use and so people will climb into the cockpit of a WWII Japanese aeroplane, say, with a FS595 fan deck to make a visual match, coming away with what they perceive as the closest colour regardless of factors like their eyesight, illumination, the age of the paint surface and the amount of chalking and/or dirt. The limitations also "persuade" observers to unconsciously select one of two biases - they either choose a colour that is close in relative brightness but not so close in hue or they focus on hue and are less sensitive to the brightness differences. Frankly whilst all these issues are important in terms of archeological colour I don't think they matter much in modelling because of the variance in applied paints, issues of scale, washes, shading and weathering, etc., which tend to balance it all out anyway. Very few people are going to measure a 1/72nd cockpit and tell you the colour is wrong but they might still say that it looks wrong. In recent years Ford went back to a colour matching system that uses a combination of human and digital assessment because they found that the purely digital system resulted in gradual error transferral over time and there was no visual assessment to say "That's wrong".

Being aware of the pitfalls, triangulating from more than one reference and being circumspect about published matches is probably good enough.

Nick

[John]

Spitfires had Marks, not marques (that's for cars,) and the designated colour lasted until the Mark 24. What we don't have is any evidence of what the actual colour was supposed to be, since no wartime shade cards have surfaced (yet.)

Or use in the certain knowledge that nobody can prove you wrong (yet.)

Or, in other words, when I state that I (and others) matched a colour, on a Spitfire, to a particular shade card from Humbrol, I'm (we're) either lying or colour-blind.

Matching actual paints is all very useful (and it is,) but, to my mind, what's supposed to be in the tin is just as important (maybe even more so.) If the paint doesn't match the shade card, but something else, it's a hint that the operative, wielding the mixer, got it wrong, not that Supermarine/Hawker/Gloster/etc. painted their aircraft the "wrong" colour.

I've related this, before, and had it poo-pooed by experts; about 30 years ago, during a visit to the Humbrol factory, I bemoaned the loss of the Authentics, and was told it was due to the then foreman checking successive batches against samples taken from the previous one, rather than the shade card. This (especially on the most popular colours) had led to colour drift, which had got so bad, it was deemed cheaper to drop the whole range, rather than call in the entire stock of tinlets, and redo them.

I never thought to ask if the same had happened to other non-Authentic colours, but it cured me of ever using paint in the tin as a guide, and taught me to look at shade cards first.

Edgar

But don't you see that the paint in the tin is the only common point of reference we have?

Nobody - and certainly not me - doubts for one minute that you matched the colour of the Spitfire parts to the chip in the chart, why on earth should anybody doubt that? The problem is that no-one else in this discussion has seen the chip you used. What we *do* have in common is access to the the actual paint that the chip claims to match, Humbrol 90. We also know, from our own experience gained over the years, that Humbrol 90 is Sky. Ergo, when you tell us that the parts you examined matched a chip of Humbrol 90, it is hardly unreasonable for the modeller to surmise that the parts were painted Sky or something very like it.

That's why I suggested earlier that the chip in the Colour System may not have been representative of the colour it claimed to match. That's not a criticism of you or an allegation of malpractice on the part of Humbrol. Saying that the parts matched a chip of Humbrol 90 in the Colour System folder, but that the chip didn't look like Sky, renders the chip useless as a means of communicating the observed colour, which is after all what we're attempting to achieve here.

John

[John]

Ah so THATS Humbrol Eau De-Nil is it? Looks like Duck Egg Green to me............

Humbrol Eau-de-Nil:

http://www.modelkitcollecting.com/topic2505.html

John

[Edgar]

But don't you see that the paint in the tin is the only common point of reference we have?

Which is hardly my fault, if I spend my time collecting information from various sources, and others don't.

Nobody - and certainly not me - doubts for one minute that you matched the colour of the Spitfire parts to the chip in the chart, why on earth should anybody doubt that? The problem is that no-one else in this discussion has seen the chip you used. What we *do* have in common is access to the the actual paint that the chip claims to match, Humbrol 90.

Sorry, but the fact you (and others) do not have a copy of a folder, issued by Humbrol, for use with their paints, cannot be allowed to invalidate it.

We also know, from our own experience gained over the years, that Humbrol 90 is Sky.

No, you don't; all you know is that paint, in certain tins, is a match for Sky; you have absolutely no evidence that 90 was Sky, nor that it was supposed to be, especially as Hornby had to change the formula to make it a match for Sky (something you consistently ignore.)

Ergo, when you tell us that the parts you examined matched a chip of Humbrol 90, it is hardly unreasonable for the modeller to surmise that the parts were painted Sky or something very like it.

Well, it might be, if there were not two Sky chips in the same file, which the parts didn't match, neither did they match my official 1971 colour card, for Sky, issued by the British Standards Institution.

That's why I suggested earlier that the chip in the Colour System may not have been representative of the colour it claimed to match. That's not a criticism of you or an allegation of malpractice on the part of Humbrol. Saying that the parts matched a chip of Humbrol 90 in the Colour System folder, but that the chip didn't look like Sky, renders the chip useless as a means of communicating the observed colour, which is after all what we're attempting to achieve here.

You might see it that way, but, having seen the actual colours, on an actual airframe, I know that, if I mix a colour that matches the chip, I will have a colour that matches the actual paint on an actual Spitfire airframe. I also know that, if I mix, or buy, a colour that is a match for Sky, it will not be a match for a colour used by Supermarine on their aircraft, because it cannot match.

[John]

"But don't you see that the paint in the tin is the only common point of reference we have?"

Which is hardly my fault, if I spend my time collecting information from various sources, and others don't.

Indeed. However the point remains.

"Nobody - and certainly not me - doubts for one minute that you matched the colour of the Spitfire parts to the chip in the chart, why on earth should anybody doubt that? The problem is that no-one else in this discussion has seen the chip you used. What we *do* have in common is access to the the actual paint that the chip claims to match, Humbrol 90."

Sorry, but the fact you (and others) do not have a copy of a folder, issued by Humbrol, for use with their paints, cannot be allowed to invalidate it.

No, but unfortunately neither does it allow us to validate it.

"We also know, from our own experience gained over the years, that Humbrol 90 is Sky."

No, you don't; all you know is that paint, in certain tins, is a match for Sky; you have absolutely no evidence that 90 was Sky, nor that it was supposed to be

Yes, we do know that 90 was meant to be Sky, This is the oldest Humbrol-issues Airfix kit in my collection, a 1986 Meteor:

Note the Humbrol colour to replace Airfix M8, post No.33 illustrates a sample of Airfix M8:

This is the painting instruction from the HiTech release of the 1/48th scale Spitfire Vb, 1992. Note the recommended underside colour:

This is the painting instruction from the Airfix rebox of the Otaki Hellcat, 1999. Note the recommended underside colour:

This is the painting instruction from the 1/48th scale Seafire III, 2002. Note the recommended underside colour:

The reference is consistent over nearly 20 years. Also, the number and dates of the tins of paints examined form, at least in my opinion, a perfectly representative sample.

especially as Hornby had to change the formula to make it a match for Sky (something you consistently ignore.)

I would respectfully refer you to post No.3, post No.10, post No.34 and post No. 40 in this thread were, far from ignoring your comment about Hornby reformulating 90, I discussed the reformulation with you. In particular in post No. 3 I made the observation:

"Whatever formulation changes Hornby may have made to Humbrol 90, the actual colour has remained remarkably consistent from it's introduction in the late 1970s until now."

"Ergo, when you tell us that the parts you examined matched a chip of Humbrol 90, it is hardly unreasonable for the modeller to surmise that the parts were painted Sky or something very like it."

Well, it might be, if there were not two Sky chips in the same file, which the parts didn't match, neither did they match my official 1971 colour card, for Sky, issued by the British Standards Institution.

That's pretty much the crux of this part of the discussion.

"That's why I suggested earlier that the chip in the Colour System may not have been representative of the colour it claimed to match. That's not a criticism of you or an allegation of malpractice on the part of Humbrol. Saying that the parts matched a chip of Humbrol 90 in the Colour System folder, but that the chip didn't look like Sky, renders the chip useless as a means of communicating the observed colour, which is after all what we're attempting to achieve here."

You might see it that way, but, having seen the actual colours, on an actual airframe, I know that, if I mix a colour that matches the chip, I will have a colour that matches the actual paint on an actual Spitfire airframe. I also know that, if I mix, or buy, a colour that is a match for Sky, it will not be a match for a colour used by Supermarine on their aircraft, because it cannot match

Yes, that's what I said.

John

[Shar2]

I think the question has pretty much been answered.