Loco model colours measured

[tomparryharry]

It's all very well having the fancy doo-dahs & nick-nacks, but is the colorimeter made in Swindon?  You can't have any old green, y'know!  Joking aside, perhaps it's wise to establish the RAL colour, and fix it as a standard.  At least that way, you 'should' know which shade you're getting.  Anywhere on the RAL 6000 spectrum should cover it.

 

Now, where's my guide dog? Here, here, doggie.... 

[Harlequin]

There are lots of RAL colours in the area but I don't think you can fix one as a standard.

 

For interest here are a bunch of GWR loco greens collected from the web, including some RAL colours (and along the bottom row a few from photos - very dangerous):

 

WARNING: This graphic almost pointless!

 

None of the colours above are specified in CIELAB colour space. The point about the CIELAB colour standard is that it takes the light source into account and so a colour is completely defined and we can calculate what it will look like under other light sources (The Johnster).

 

There are several different RAL standards (just to make life even more confusing) but only one of them, RAL Design, has a well-defined mapping to CIELAB space. (None of the RAL colours above are specified in RAL Design.)

 

[Nearholmer]

Given all of the variables, I think it best to pick the colour you like, and use that. possibly not accurate, but certainly happy!
 

For me, the deeply saturated middle chrome green recommended for pre-1928 by GWR.org, which seems to convey the conservative dignity of the GWR, whether it be ‘correct’ or not, and for BR-period the Humbrol Brunswick Green, because I got tiny spots of it on my gran’s dining table when painting an Airfix ‘Evening Star’ when there were still steam locos in traffic on BR - I thought she wouldn’t notice, but .......

[The Stationmaster]

One thing to remember is that it is along time since 4073 (in particular) and 2516 were last repainted and until a year or so ago they were still being regularly cleaned by former Enginemen using a traditional paraffin and oil mixture so the colours they have now are quite an age and ageing on from when they were repainted in Swindon Works to go into preservation.

 

It would be interesting to see what the GWS use and the SVR

[Harlequin]

I'm struggling to get to grips with the results from the two colorimeters...

 

You would think that devices like these, which claim to be calibrated, would give consistent results but I'm finding that they differ from each other and worse, not in a consistent way.

 

So at the moment I don't trust either of them!

 

Edited October 4, 2020 by Harlequin

[tomparryharry]

The problem with some hand-held devices can be open to question. An example of a lamp-post doing 81 MPH in a 30 zone. The worst part of this is because the authorities use the word 'computer', convictions can be obtained in somewhat dodgy circumstances. 

 

With respects to our police officers, who are by & large, only human.

[Harlequin]

To assess which of my two devices is the more accurate I needed a well defined set of colours that I could measure so I bought a CIELAB colour fan:

 

 

Here are a typical set of results for the colour in the middle of the screen CIELAB = 75 -20 50:

 

You can see that the result from the Chinese device on the left is significantly further away (76.18, -13.08, 37.97) than the Canadian Nix device on the right (76, -21, 49). Similar results are seen when measuring random colours across the fan - the Chinese device is never as close as the Nix.

 

So I will be trusting the Nix from hereon in. The Nix uses a D50 illuminant - this is important to know because it allows comparison of colours independently of light sources.

 

Just as an initial indication, the first data point: The Hornby Prairie 6110 measures as CIELAB = 24 -6 8 (right tank),  or CIELAB = 25 -5 7 (left tank).

 

More data to follow...

 

Edited November 8, 2020 by Harlequin

[FoxUnpopuli]

On 04/10/2020 at 11:22, Harlequin said:

WARNING: This graphic almost pointless!

 

 

 

But it's so pretty!

 

37 minutes ago, Harlequin said:

You can see that the result from the Chinese device on the left is significantly further away (76.18, -3.08, 37.97) than the Canadian Nix device on the right (76, -21, 49). Similar results are seen when measuring random colours across the fan - the Chinese device is never as close as the Nix.

 

 

eBay awaits.  People will buy anything these days.  We bought a car maybe four years ago for £700, running, with MOT.  Just sold it as MOT fail with some recent unfortunate damage for over a grand. 

 

Down is up I swear.

 

I digress.  I have to agree with an opinion above - pick a paint, if it looks good, it's right.  I would also suggest that pretty much all the 'colours' in your lovely chart could be observed on GW locos at one time or another, so maybe you can't get it wrong!

Edited November 8, 2020 by FoxUnpopuli
Clarfication of bad grammar. (Bad Grammar! Back to bed!)

[Butler Henderson]

This review https://www.cnet.com/news/3-color-sensor-gadgets-to-take-the-pain-out-of-paint-matching/ describes the Nix device as being the best, but reading the review that probably should be the best of bad bunch as it plainly never gets RGBs spot on, rather it generally gives nearer readings than the majority  given by the other two devices in the review.

[Harlequin]

10 hours ago, Butler Henderson said:

This review https://www.cnet.com/news/3-color-sensor-gadgets-to-take-the-pain-out-of-paint-matching/ describes the Nix device as being the best, but reading the review that probably should be the best of bad bunch as it plainly never gets RGBs spot on, rather it generally gives nearer readings than the majority  given by the other two devices in the review.

Thanks, that’s really interesting but the reviewer in this article has not understood that you can only compare colours numerically if you know what light source was used. She looks up the manufacturer’s stated RGB value of a paint colour then measures a painted sample. Apart from RGB not including illuminant info, there are too many uncontrolled variables in that process. Having carefully explained how DeltaE is used to describe colour differences she doesn’t state the DeltaE between the hoped for RGB values and the measured values - and of course she can’t because she’s not working in CIELAB space with known illuminants.

 

Edited November 9, 2020 by Harlequin

[Il Grifone]

On 28/09/2020 at 20:25, Butler Henderson said:

Always struck me the one opportunity that no model paint supplier has taken is to offer a range that matches RTR models with appropriate descriptions to a overcome brand issues like Blue Branch Line BR Green and Red Box LSWR 488 Green.

 

I can see two problems here: copyright and too much hassle to be worthwhile commercially. (Not that it's not a good idea.)

[spamcan61]

2 minutes ago, Il Grifone said:

 

I can see two problems here: copyright and too much hassle to be worthwhile commercially. (Not that it's not a good idea.)

I don't see a problem unless the colour has been registered as a trademark, which only works in the USA  (e.g. UPS brown)  IIRC; agree it's a good idea in principle but too much faff, especially given the manufacturer's own variation in colours - thinking of my Bachmann Mk1 coaches in various shades of green for example.

[Il Grifone]

Copyright is a touchy matter (like anything where lawyers are involved - e.g. U.S. elections!). I've just tried to look up Humbrol colour charts and kept getting "Image removed at copyright holders request"!

 

Interestingly the colour charts agree with my perception of the GWR colours, but not of the Humbrol Brunswick green. Back in the day, I used this as BR/GWR green (it was cheaper than their 'GWR loco green' - 7d rather than 1/3 IIRC). I touched up a chip on the tender of my HD 'Montrose' with it and recall it as a tad darker but near enough.

Edited November 11, 2020 by Il Grifone

[Butler Henderson]

Whatever the accuracy or not of the GW Green applied on Airfix GMR items was, Tamiya Japanese Army green is a good match.

[JimC]

On 28/09/2020 at 19:25, Butler Henderson said:

a range that matches RTR models with 

That presupposes that every RTR model is in fact exactly the same colour. That seems to me a pretty big assumption. If every example of a short run model is moulded from the same batch of plastic then sure, but it seems to me, with some (admittedly 40 years ago ) experience in plastics manufacture that anything that's produced over a period of time with different batches of plastic or paint is liable to vary in shade. I still recall with a small shudder the feeling of a colour match going wrong, and the struggle to get something that might do before the batch goes over temperature and has to be abandoned.