British Aviation Colours of World War Two book insert - Spectrophtometer readouts and sample recipes using Tamiya, Liquitex and Golden Fluid Acrylics

[Steben]

Yellow ochre and sienna are very closely related.
The main difference between the two is a portion manganese oxide, which turns the sienna darker than ochre.
"Yellow ochre" is a very diverse family with its combinations of FeO-OH (yellow oxide) and Fe2o3 (red oxide).
Every "burnt" pigment turns FeO-OH into Fe2o3, making it redder.
But other oxides are present sometimes adding some variation.

The reason I come up with this time after time is the enigma around "PC10" which is often stated to be "ochre and black" based on some recipes, just as is done with olive drab.
Pure yellow iron oxide and black gives olive green. Pure red iron oxide and black gives maroon. All mixes of both (present in different natural ochres) will yield a olive drabbish or brownish drab.
That's why PC10 is not olive green at all, since the "ochre" proscribed always had a big portion Fe2o3 in it. It's all about a dark orange. PC10 can be mimicked with yellow oxide, some red oxide and black.
And why US ww2 olive drab is much greener, since a classic pigment recipe calls for specific yellow iron oxide (not random "ochre"!) and even some chrome yellow, which is not even an earthy colour.

Edited May 12, 2022 by Steben

[Casey]

28 minutes ago, Steben said:

Yellow ochre and sienna are very closely related.
The main difference between the two is a portion manganese oxide, which turns the sienna darker than ochre.
"Yellow ochre" is a very diverse family with its combinations of FeO-OH (yellow oxide) and Fe2o3 (red oxide).
Every "burnt" pigment turns FeO-OH into Fe2o3, making it redder.
But other oxides are present sometimes adding some variation.

The reason I come up with this time after time is the enigma around "PC10" which is often stated to be "ochre and black" based on some recipes, just as is done with olive drab.
Pure yellow iron oxide and black gives olive green. Pure red iron oxide and black gives maroon. All mixes of both (present in different natural ochres) will yield a olive drabbish or brownish drab.
That's why PC10 is not olive green at all, since the "ochre" proscribed always had a big portion Fe2o3 in it. It's all about a dark orange. PC10 can be mimicked with yellow oxide, some red oxide and black.
And why US ww2 olive drab is much greener, since a classic pigment recipe calls for specific yellow iron oxide (not random "ochre"!) and even some chrome yellow, which is not even an earthy colour.

That explains why my math does call for purer yellow.

 

I always wonder if chrome yellow or similar class of pigment was not added there for its toxicity more than a color... it had been replaced by cadmium yellow - which I don't use neither since it is also having some toxicity concerns (and having an Prop65 label in California for example)

[Ngantek]

Tomorrow if I find the time, I was going to make a comparison of various iterations of EDSG and sky (a few tamiya mixes including these, humbrol, both types of gunze), so I'll post them here if that's helpful to anyone. Photos in sunlight I guess? any other suggestions (I'm a colour blind oaf with a phone camera on auto setting) that would make it useful? Sadly I don't have colour chips which would probably be necessary for it to mean much.

[Casey]

1 hour ago, Ngantek said:

Would it be really annoying (or even worthwhile) to try some of the more difficult mixes by maybe starting with an extra colour most adjacent to the final match. So xf52 for dark earth say. So each one would the 'closest match' colour with the algorithm matching using 1 to 14?

My algorithm would do that if I had paint data for xf52 I only have measured XF1-XF16 so far, this is time consuming and expensive process.

[Ngantek]

Not particularly useful to this but since I did it anyway...

These are all airbrushed onto styrene, which has been primed with Gunze Mr surfacer grey.

 

Here's various versions of sky. From the wingtip in, 

-Humbrol 90 enamel (sky)

-Tamiya xf21 (sky), 1 part :  xf2 (white) 1 part

-Your mix xf2 (white) 28 : xf14 (JA grey) 21 : xf5 (green) 1

-Plain xf21 (sky)

-Gunze c368

-Gunze H74

 

The  top photo is in direct sunlight. The bottom outside in the shade on a sunny day.

 

 

 

 

For EDSG we have, again wingtip to centre

-Humbrol 123 enamel

-bare primer

-'Roy Sutherland' approximation. Tamiya xf24 (dark grey) 17 : xf2 (white) 2 : xf8 (blue) 1

-Your mix, xf2 (white) 16 : xf1 (black) 4 : xf8 (blue) 2 : xf7 (red) 1

-Gunze c333

-Gunze h333

Again top is direct sun, bottom is shade.

 

 

 

 

 

Just a quick thing @Casey you've mixed the labels for black and white  (black = xf1, white = xf2).

 

These were all measure by mass rather than volume.

 

In both cases, relative large ratio between major and minor components means that, given I was only mixing a gram or so, there may  minor inaccuracies.

 

 

 

 

[Steben]

I see minor differences in gloss. This may disrupt impression. An overall varnish might equalise but won't take away all differences.

Edited May 13, 2022 by Steben

[Scimitar F1]

3 hours ago, Ngantek said:

-Your mix xf2 (white) 28 : xf14 (JA grey) 21 : xf5 (green) 1

To my eye this looks pretty much on the nail

[Casey]

3 hours ago, Ngantek said:

Not particularly useful to this but since I did it anyway...

Actually it is useful. I do miss the comparison with the reference on your photo, but I do have H90, XF21, H123 and reference chips, so I can reproduce part of your image.

 

Also thanks for noting the mislabel. I did really switch those two. White is a very special paint for me, since it is a source of scatterer for calculations and in *all* sets it is numbered as 1 Label is correct.

 

Let me start with theorycraft first

 

Here are theoretical calculation results for those two paints:

Now let me make those measure their spectra plus photo them with original source.

 

Stay tuned, watching the paint dry!

 

Edited May 13, 2022 by Casey

[Ngantek]

19 minutes ago, Scimitar F1 said:

To my eye this looks pretty much on the nail

I'm acutely aware of the folly of a colour blind person such as myself commenting on the quality of colour matches!

 

But yeah it looks the best with the naked eye, at least what I imagine as sky. Though you see a more washed out colour in some photos.

 

As to the EDSG, they all look like recognisable variations. @Casey match is a good 'faded blueish' version one might see on, say the duxford sea vixen. Gunze reminds me of fresh painted SHAR and humbrol the very indistinct temperate sea scheme with Dslate grey. It looks rather muddy though.

[Ngantek]

2 hours ago, Steben said:

I see minor differences in gloss. This may disrupt impression. An overall varnish might equalise but won't take away all differences.

Yeah that's sensible. I'll give it a layer next the varnish is out.

[Casey]

1 hour ago, Ngantek said:

I'm acutely aware of the folly of a colour blind person such as myself commenting on the quality of colour matches!

I wonder if actually a spectrophotometer would be an useful tool (think: glasses are also a device to help with eyesight...) for you.

 

I found a couple of them on Amazon and some cheaper one go down to even as low as 65$ (https://www.amazon.com/dp/B09QBYD5VW/), with mid range AMT599 for 300$ (https://www.amazon.com/dp/B07Q5FC8WT/).

 

Professional ones do go for thousands of dollars, but depending on a budget and if you mix a lot of paints, the cheap range one can be an useful investment.

 

The cheap one is directional meter with 45/0 geometry, AMT is diffuse light d/8 SCI

 

Here is more throrough explanation - https://blog.hunterlab.com/blog/color-and-appearance-theory/instrument-geometry-directional-450-or-diffuse-d8-sphere/ but generally: 45/0 mimic the human eye and is equally sensitive to gloss level (also are smaller and cheaper), while diffuse spectrophotometers can ignore the gloss level (and paint texture). More expensive ones can measure both reflected and diffuse light at one go (those have SCI and SCE mode switch or do two measurements per sample)

 

Edited May 13, 2022 by Casey

[Casey]

@Ngantek, @Scimitar F1 @Steben

 

Here are the results:

 

 

This is an scan of real samples vs my mixes and comparison to the alternatives I had.

 

I copy&pasted part of the image in paint to show them visually closer.

 

And here are the spectrophotometer data:

RAF0009 - Extra Dark Sea Grey

My match is closer than Humbrol there.

 

 

RAF0014 - Sky

My mix is the the closest match there too

 

The measured weight ratios were:

 

EDSG: 16.08 : 3.99 : 2.19 : 1

Sky: 28 : 21.39 : 1

 

[Scimitar F1]

On 5/13/2022 at 2:47 PM, Ngantek said:

Your mix, xf2 (white) 16 : xf1 (black) 4 : xf8 (blue) 2 : xf7 (red) 1

The EDSG is more interesting. The sample mix here is much more of an outlier than the Sky one. Most of us will only have been exposed to EDSG on FAA aircraft post-war as there are so few wartime shots of Coastal Command aircraft in colour and the I don’t think the films were particularly accurate. 

 

For almost every shot of an FAA aircraft (and I am partial - check out my username!) the EDSG appears much lighter than the model paints of the same colour. Every time this is mentioned fading is immediately wheeled out as the reason for the disparity. Seeing this thread I am not convinced by the latter argument any longer - particularly as carrier borne aircraft are constantly retouched for corrosion and so should look like they had the pox but do not in most photos. I think that this may be the explanation!

 

@Casey please keep up the great work. You have moved the argument on more in this thread than I have seen. Colour is a science and not an art in as much as it is measurable and repeatable - your technical approach has merit!

Edited May 30, 2022 by Scimitar F1

[Steben]

Excellent work.

Such a shame paint companies tend to change their mixes from time to time.
This happens here and there now and then.

[Casey]

1 hour ago, Scimitar F1 said:

The EDSG is more interesting. The sample mix here is mush more of an outlier than the Sky one.

 

Almost every shot of an FAA aircraft (and I am partial - check out my username!) appears much lighter than the paints of the same colour.

Does it mean my mix (and the book reference) is actually more accurate?

Long time ago @Nick Millman wrote:

Quote

 

FS 26118 is closest to EDSG @ 0.84 (where less than 2.0 = a close match) so a very close match. FS 36118 is @ 4.61 and lighter in appearance. This difference is caused by the semi-gloss vs matt finish of the FS 595b values.

FS 36118 is closest to DSG @ 2.40 but is darker. FS 36173 as suggested by Geoff Thomas is @ 6.56 and too light.

Engine Gray to EDSG is @ 6.76 (different hues!)

Munsell values:-

EDSG 1.9 PB 3.4/1.1

DSG 0.2 PB 4.2/1.2

Same hue but different degrees of lightness and saturation as might be expected.

These comparisons were made by measuring the MAP swatches with a spectrophotometer to obtain their L*a*b* values and then comparing those to the FS595b values using the DE2000 formula recommended by the Commission International de l’Éclairage (CIE).

 

 

Quick link to the measured databases show:

 

FS26118: L41.85 a-1.07 b-5.22

My mix: L41.51 a-1.72 b-5.54

 

https://www.e-paint.co.uk/lab-hlc-rgb-lrv-values.asp?cRange=Federal+Standard+595C&cRef=26118&cDescription=Gray+#11+/+Medium+gunship+gray

 

Let me see if I have FS26118 real chip laying around...

 

Update: Nope. I only have 1950's 595C, FS26118 was added in 1989.

Edited May 14, 2022 by Casey
Not having FS26118...

[Steben]

RAL 7031 looks promising according to e-paint. Needs a touch of black or a little bigger touch of RAL 7021.
When compared to my RAF chips though, it looks an eyeball match to ocean grey. And EDSG much darker.

Edited May 14, 2022 by Steben

[Casey]

Just now, Steben said:

Excellent work.

Such a shame paint companies tend to change their mixes from time to time.
This happens here and there now and then.

Pigments are not constant, they vary from batch to batch, so it is normal to have to change the recipes or do per-batch adjustments.  Some pigments become unavailable, some become forbidden. Some small revolution about cadmium yellow ban happens...

 

It is kind of a moving target.

[Steben]

Left RAL 7015 (EDSG), right RAL 7031 (OG)


https://scontent-bru2-1.xx.fbcdn.net/v/t1.15752-9/280807438_831646464904990_2862855817900477730_n.jpg?stp=dst-jpg_s2048x2048&_nc_cat=103&ccb=1-6&_nc_sid=ae9488&_nc_ohc=LsjDDM56SJ4AX-svWou&_nc_ht=scontent-bru2-1.xx&oh=03_AVJRreX56infVw6aIl_u0ymrdKuLFb2zZ2QlOHrbsYhmSQ&oe=62A50EF3

Edited May 14, 2022 by Steben

[Casey]

Digging through http://aviationarchives.net/FederalStandard595C.pdf

 

FS26118 has following ingredients:

• Rutile Titanium Dioxide (XF2 is titanium dioxide)
• Phthalo Blue (Green Shade)
• Carbazole Violet (it is dioxazine purple, and as far as I know Pigment Violet 23 (PV23) was not available as a paint before 1953 which means it may have been unlikely to be used in WWII...)
• Carbon Black (Blue Shade) (XF1 is carbon black)

Edited May 14, 2022 by Casey

[Casey]

14 minutes ago, Steben said:

Left RAL 7015 (EDSG), right RAL 7031 (OG)


https://scontent-bru2-1.xx.fbcdn.net/v/t1.15752-9/280807438_831646464904990_2862855817900477730_n.jpg?stp=dst-jpg_s2048x2048&_nc_cat=103&ccb=1-6&_nc_sid=ae9488&_nc_ohc=LsjDDM56SJ4AX-svWou&_nc_ht=scontent-bru2-1.xx&oh=03_AVJRreX56infVw6aIl_u0ymrdKuLFb2zZ2QlOHrbsYhmSQ&oe=62A50EF3

This data may help you a bit

[Steben]

Thx Casey.

Again, an eyeball match just within dE < 2.00 🤣
Just thinking, standard RAL is satin reference. A flat one will be slightly lighter, perhaps even closer.

Can you put RAL7031 against ocean grey?

Edited May 14, 2022 by Steben

[Troy Smith]

14 hours ago, Casey said:

Sky: 28 : 21.39 : 1

 

Regarding Sky.  I found XF-21 too 'muddy'  (there is a slight brown to it)  and no amount of white will cure that. 

 

I used X-23 Clear Blue, it looks to be ultramarine,  with XF-59 Desert Yellow, which is a yellow ochre,  x-23 :1 to XF-59 : 3, and then add lots of white. Can't remember how much, but lots,   I think I used 0.1ml X-23, 0.3ml XF-59 and 1 ml white, maybe 1.5ml

  

One here somewhere @Nick Millman  had mentioned the pigments in Sky were ultramarine and yellow ochre, which is why i experimented with those (as well as various others)  

This gave a the correct very pale yellow green appearance.  

 

The various Tamiya clear colours maybe worth playing with, while they are in a gloss carrier,  they are quite intense pigments,  so as the amounts needed are small for 'shifting' the hue in a mix.

 

@Steben regarding EDSG and Ocean Grey,  note that the mixes above for the Sea Grey have all included blue and red,  which confirms that the Sea Grey family have subtle purple hue, when I was working on Medium Sea Grey,  I used Tamiya XF-82 Ocean Grey,   which is blue hued.

MSG used XF-82 plus white, and a tiny amount of red, which got me a good visual match for the chip.

from memory it was 3ml XF82, 3.5ml white and 0.1 ml red.

 

Ocean Grey  is noted as having a slight green hue,  again, I used XF-82, with a little white and a tiny amount of yellow.

 

I have got the mixes noted down, and did then make a small samples using a 1ml syringe.  but I'd need to go on a dig to find them,  which might take  minute, or turn into an hour of tidying up the chaos.....  and did then make a small samples using a 1ml syringe.

 

FWIW, I have laid brush out chips over the RAF museum chips in direct sunlight and I can't see a difference, and that is a very unforgiving light source.

 

On 12/05/2022 at 21:01, Casey said:

My algorithm would do that if I had paint data for xf52 I only have measured XF1-XF16 so far, this is time consuming and expensive process.

I can send you brush out of Tamiya colours you don't have if that helps, along with brush out of my mixes, be interesting to see how they scan.

 

HTH

 

 

[Casey]

7 hours ago, Troy Smith said:

I can send you brush out of Tamiya colours you don't have if that helps, along with brush out of my mixes, be interesting to see how they scan.

It is not the lack of paints that is limiting me there, but the silly fact that Earth is spinning and only giving me 24 hours per day, more or less.... I have all of Tamiya acrylics range in multiple batches, couple of liters of them total.

[Rolls-Royce]

@Casey it appears by your data fields that you may be using BabelColor's PatchTool or something similar. I just started with that myself in conjunction with their CT&A software and an XRite I1 Pro 2 (and now a Pro 3 Plus). 

Edited May 30, 2022 by Rolls-Royce

[Casey]

11 minutes ago, Rolls-Royce said:

@Casey it appears by your data fields that you may be using BabelColor's PatchTool or something similar. I just started with that myself in conjunction with their CT&A software and an XRite I1 Pro 2 (and now a Pro 3 Plus). 

I use Matlab and my own math implementations. The data fields are de facto standard for all color analysis software so they should be same.