Colourisation: problems ahead.

[adb968008]

Theres a bigger problem than colour coming…

 

heres my LMS 8f 2-8-0 in blue…

 

 

this tech is only getting smarter, before we know locomotives that never were will have a whole new meaning.

Edited March 14 by adb968008

[melmerby]

30 minutes ago, Wheatley said:

a few brown and cream ones. 

All corridor coaches?

IIRC the suburban ones were always some shade of red on the WR.

[Wheatley]

1 hour ago, melmerby said:

All corridor coaches?

IIRC the suburban ones were always some shade of red on the WR.

All corridor, the original intention was (I think) that they were for named trains. About 300 miles south of my area of interest though. 

[BachelorBoy]

13 hours ago, Wheatley said:

80/20 Rule,

 

 

How do you apply that to modelling, please?

 

[BachelorBoy]

Anyone fancy identifying this poorly colourised B+W image?

 

[Wheatley]

3 hours ago, BachelorBoy said:

How do you apply that to modelling, please?

 

 

1. 80% of perfect gets you a respectable result. The remaining 20% takes 80% of the total time for the project, whatever it is. 

 

Practical examples - modelling in OO instead of EM or P4, missing out all the bits of brake gear you can't see unless you turn the wagon upside down, basing a model on a single photo and accepting you might have misinterpreted it. 

 

3 hours ago, BachelorBoy said:

Anyone fancy identifying this poorly colourised B+W image?

 

 

It's the Juvinus Molestus Sorcero Cashcowium Rapido.  I could do some proper research into what the Latin  for "Annoying Teenage Wizard Express" is but that'll do for the purpose in hand.  It should be in Rowling's Vastly Improved Engine Green btw. 

 

I did actually accept that this would be a problem for hobbyists because our research tends to be more casual and we accept things at face value which a 'proper' researcher wouldn't. We also increasingly just ask other people and accept the answers instead of putting a bit of effort in, there are several threads on here along the lines of "Can anyone tell me what sort of trains were operating on this area in this era ?". I don't know, maybe go and look at a few photos ?   It also becomes a self-feeding problem because the Internet Experten gradually achieve a largely undeserved reputation when in fact a significant minority of them are taking tosh most of the time.

 

It's not exclusively a railway thing. If you want to start a fight in an IPMS group just ask them what colour Olive Drab should fade to. There's even a sort of Godwin's Law about that particular question as to how long it will take for someone to post that photo of a B17 with every panel a different colour.  

 

 

 

Edited March 15 by Wheatley

[BachelorBoy]

2 hours ago, Wheatley said:

It's the Juvinus Molestus Sorcero Cashcowium Rapido.  I could do some proper research into what the Latin  for "Annoying Teenage Wizard Express" is but that'll do for the purpose in hand.  It should be in Rowling's Vastly Improved Engine Green btw. 

 

A good guess!

 

The actual prompt for the AI program (Midjourney) was:

 

"1940s steam locomotive designed by Oliver Bulleid. Old black and white photo style"

 

I think Midjourney needs a lot more work.

 

 

 

 

[Wheatley]

On that evidence I don't think it's going to be an issue for a while !

 

[BachelorBoy]

7 minutes ago, Wheatley said:

On that evidence I don't think it's going to be an issue for a while !

 

Sorry this is taking the thread off-topic a bit,  but after that I gave the Midjourney the prompt "Oliver Bulleid" by itself (no mention of steam locos at all) and this is what it came up with.  

 

[Captain Slough]

the first one with the obviously GWR engine has MK1 coaches with 5 windows between each door set.....   75ft chassis version? AI is weird.

[ejstubbs]

On 14/03/2024 at 06:49, Wheatley said:

For further examples see the Victorian interpretation of an inguanadon with the thumbspike on its nose (its in a park somewhere)

 

It's in Crystal Palace Park - there's actually two of them.  The mistake was made by Gideon Mantell, the discoverer of the first, incomplete, iguanodon fossils.  In his defence, he was working with the specimens he had at the time; it was only when more complete specimens were found later that the mistake was recognised.  Bear in mind that he made his first skeletal reconstruction in 1834, and the Crystal Palace iguanodons, which were built nearly 20 years later, were still constructed with the erroneous nose spike or horn which they retain to this day*.

 

What's probably more egregiously wrong about the Crystal Palace iguanodons is that they are depicted as heavy, pachyderm-like creatures, contrary to what Mantell had worked out about them five years earlier.  This error was based on the views of another paleontologist Richard Owen, who still clung to creationist ideas and believed that the iguanodon was fundamentally mammalian, and could not have "transmuted" from a reptilian form into modern mammal-like species.   (Owen became the Superintendent of the natural history departments of the British Museum in 1856, which gained their own premises in 1880 in what we now know as the Natural History Museum.  His rather forbidding statue stood at the midway point of the main staircase until 2009, when it was replaced by a statue of Charles Darwin having a nice sit down - but no cup of tea 🙁.  Owen was widely regarded as being not a very nice person, to put it mildly.)

 

* They were retained during the 2001 renovation programme, thankfully, preserving the original mistake.  All the models in the Crystal Palace Dinosaur Park were kept as as close to original condition during the renovations as was practically possible, given the state of deterioration of some of them.  Some had actually gone missing, and had to replaced with fibre-glass replicas.  As of 2007 the site is Grade I listed, and quite right too.  When I was a nipper and my family lived in Bromley, we used to visit the dinosaurs quite regularly.  They were looking pretty care-worn even back then.  I was very pleased to find them lovingly restored when I visited the park again in the early 2000s, more than 40 years after I'd last been there, despite it being a bitingly cold day (we eventually retired to the nearby indoor cafe for a restorative cuppa and fish finger sandwiches - the latter being a delicacy I had not previously experienced but which proved to be eminently sustaining in such weather)..

[Mike Buckner]

On 12/03/2024 at 09:44, LBRJ said:

There is also the point that everyone pretty much sees colours differently anyway, at least to some degree.

 

This is certainly true. My own two eyes see things slightly differently. If I close my right eye, and view with my left, things look slightly more pink, than if I close my left eye, and view with my right, when they look slightly more blue.

 

[pete_mcfarlane]

The Crystal Palace dinosaurs are really important, because they show how the scientific consensus evolves in the light of more evidence. 

 

16 minutes ago, ejstubbs said:

This error was based on the views of another paleontologist Richard Owen, who still clung to creationist ideas and believed that the iguanodon was fundamentally mammalian, and could not have "transmuted" from a reptilian form into modern mammal-like species.

The irony is that we now think they are related to birds. As for Owen, he seems to be up there with John Chester Craven in the Pantheon of unpleasant Victorians. Nowadays we probably recognise him as having some kind of personality disorder.

 

Anyway, as for colourisations, the best story I heard was about the colourised version of King Kong. Having gone to great lengths to do this, Fay Wray pointed out that they'd got the colour of her dress wrong. 

[BachelorBoy]

1 minute ago, pete_mcfarlane said:

The Crystal Palace dinosaurs are really important, because they show how the scientific consensus evolves in the light of more evidence. 

 

 

A doctor won the Nobel prize in 1949 for his work on lobotomy,

 

 

[melmerby]

2 hours ago, BachelorBoy said:

 

A doctor won the Nobel prize in 1949 for his work on lobotomy,

 

 

Old joke coming up:

I rather have a full bottle in front of me,

than a full frontal lobotomy......😁

[TheSignalEngineer]

On 12/03/2024 at 09:44, LBRJ said:

There is also the point that everyone pretty much sees colours differently anyway, at least to some degree.

The colour rendition of my two eyes is not an exact match.

[Artless Bodger]

8 hours ago, ejstubbs said:

 

It's in Crystal Palace Park - there's actually two of them.  The mistake was made by Gideon Mantell, the discoverer of the first, incomplete, iguanodon fossils.  In his defence, he was working with the specimens he had at the time; it was only when more complete specimens were found later that the mistake was recognised.  Bear in mind that he made his first skeletal reconstruction in 1834, and the Crystal Palace iguanodons, which were built nearly 20 years later, were still constructed with the erroneous nose spike or horn which they retain to this day*.

 

What's probably more egregiously wrong about the Crystal Palace iguanodons is that they are depicted as heavy, pachyderm-like creatures, contrary to what Mantell had worked out about them five years earlier.  This error was based on the views of another paleontologist Richard Owen, who still clung to creationist ideas and believed that the iguanodon was fundamentally mammalian, and could not have "transmuted" from a reptilian form into modern mammal-like species.   (Owen became the Superintendent of the natural history departments of the British Museum in 1856, which gained their own premises in 1880 in what we now know as the Natural History Museum.  His rather forbidding statue stood at the midway point of the main staircase until 2009, when it was replaced by a statue of Charles Darwin having a nice sit down - but no cup of tea 🙁.  Owen was widely regarded as being not a very nice person, to put it mildly.)

 

* They were retained during the 2001 renovation programme, thankfully, preserving the original mistake.  All the models in the Crystal Palace Dinosaur Park were kept as as close to original condition during the renovations as was practically possible, given the state of deterioration of some of them.  Some had actually gone missing, and had to replaced with fibre-glass replicas.  As of 2007 the site is Grade I listed, and quite right too.  When I was a nipper and my family lived in Bromley, we used to visit the dinosaurs quite regularly.  They were looking pretty care-worn even back then.  I was very pleased to find them lovingly restored when I visited the park again in the early 2000s, more than 40 years after I'd last been there, despite it being a bitingly cold day (we eventually retired to the nearby indoor cafe for a restorative cuppa and fish finger sandwiches - the latter being a delicacy I had not previously experienced but which proved to be eminently sustaining in such weather)..

OT, the first reasonably complete iguanadon skeleton was found in Benstead's quarry in Maidstone. Consequently, Maidstone is unique in having a dinosaur in its coat of arms.

 https://museum.maidstone.gov.uk/explore/collections/geology/maidstone-and-the-iguanodon/

From this website I find that there is now a sculpture of an iguanadon outside the refurbished Maidstone East station (I haven't been back to Maidstone for a few years now, but we used to walk past the quarry some Sunday afternoons in my childhood). 

[The Johnster]

Any discussion on pinpointing an exact colour is going to run and run, mostly in circles.  There are just too many variables; perception on the day the photo was taken, ambient light conditions (not just strength but cast and diffusion), reflectivity of the surface, degradation of the image in the camera lens and by the film or light sensitive panel, whatever printing method is used, the paper the image is printed on or the qualities of the monitor screen, ambient lighting in the viewing room, and ambient colour in the viewing room, plus ambient outside light if the room has windows and possible glasses tints.  Good luck with all that, gentlemen.  Oh, and your eyes see large blocks of colour differently to smaller areas.  
 

I work to the principle that of it looks like the colour I remember, that’s good enough; any more is a hiding to nothing, IMHO. 

Edited March 16 by The Johnster

[BachelorBoy]

8 hours ago, The Johnster said:

 

I work to the principle that of it looks like the colour I remember, that’s good enough; any more is a hiding to nothing, IMHO. 

 

What happens in the future when there is no-one alive who remembers the colour? This is where I think colourisation will muddy the water.

 

[Mark Saunders]

13 hours ago, BachelorBoy said:

 

What happens in the future when there is no-one alive who remembers the colour? This is where I think colourisation will muddy the water.

 

It’s bad enough when people are alive who remember the real colour but colour perception is personal and likewise memories!

[BernardTPM]

I think the problem here is not so much the exact shade and perception of colour, which is really another well discussed topic, but the possible use of random colour and guesses when converting a monochrome image. Personally I'd rather have the original non-colour picture.

[The Johnster]

In the future when nobody has direct memory of the colours, much like the present when nobody has direct memory of the colours, it is advisable for modellers to consult records of the exact paint mixes used, where they are recorded and the information is available.  Our experience of, say, LNWR coach livery or Stroudley 'Improved Engine Green' and many other liveries is informed by museum pieces in museum condition, and lighting.  What these colours looked like in service when weathered and faded a bit (we all know that pre-grouping locos and stock was always kept in immaculate condition, but steam railways are dirty places even when they are not located in heavily polluted industrialised areas) is anyone's guess, as nobody is alive who remembers them in daily use (I rather doubt the small number of surviving centenarians from pre-grouping days bothered to take much notice). 

 

I agree that colourised colours will make very poor guidance for modellers or anyone else interested in liveries, but no more so than existing paintings of pre-grouping locos and stock.  There are some examples extant that are probably pretty reliable; the NRM's collection of NER locos for example were painted at Darlington not long after the NER had ceased to exist, and the J72 pilots and J69/N17 pilots of the 1958 era were painted at Darlington and Stratford by men who may well have worked in the paint shops in pre-grouping days.  But by and large copying preserved locos and stock is fraught with all sorts of accuracy issues. 

 

We can, I contend, only do our best with whatever information we have, and be prepared to modify our models if better information comes to light, but I still contend that attempting to pin down colours with any degree of accuracy when the only reference material is colour film photographs is a pointless waste of time, and no worse than relying on colourised images. 

[kevinlms]

38 minutes ago, BernardTPM said:

I think the problem here is not so much the exact shade and perception of colour, which is really another well discussed topic, but the possible use of random colour and guesses when converting a monochrome image. Personally I'd rather have the original non-colour picture.

Looking at it another way, what happens if you take a modern digital photo (which will be in colour, because almost no one takes black & white originals now), convert it to monochrome and saves as such. What happens if you take that saved image and re-colourise it?

[BachelorBoy]

47 minutes ago, kevinlms said:

Looking at it another way, what happens if you take a modern digital photo (which will be in colour, because almost no one takes black & white originals now), convert it to monochrome and saves as such. What happens if you take that saved image and re-colourise it?

 

Original

Photoshop conversion to B+W using default settings

Photoshop colourisation using default settings

 

 

[BachelorBoy]

1 hour ago, BernardTPM said:

I think the problem here is not so much the exact shade and perception of colour, which is really another well discussed topic, but the possible use of random colour and guesses when converting a monochrome image. Personally I'd rather have the original non-colour picture.

 

You have hit the nail on the head. Thank you.