How do we perceive realism in a model railway?

[Ian Holmes]

Some random thoughts on my blog today about how we perceive realism in our models.

I don't know if there's any optometrists/opticians out there who can offer a simpler interpretation of my ramblings.

https://iansmicrolayouts.blogspot.com/2024/01/eye-level.html

 

 

Ian

[rovex]

I look at realism in two ways, there are those layouts which are accurate down to the last rivet which I admire greatly but could never aspire to, and then there are those layouts which don't attempt to that level of detail but manage to capture the atmosphere of the real thing.

The mind/eye can be very forgiving 

[Cwmtwrch]

I know very little about the human visual system, except that it responds to movement, and to its owners point of interest/concentration [which is all that is normally in focus if you are close to the model]. Realism is subjective, and very much depends on what the person involved knows about the original of the model as well as other factors such as colour, weathering, relationships between the parts of the overall scene, movement which mimics [or doesn't] the movement of the original, and our actual view of the scene modelled. As Rovex says, we can be aware of minor flaws, but able to ignore them if the overall picture convinces, within the constraints of what is possible at the scale used, but this again is entirely subjective.

 

I fear your maths is a bit awry; in OO a model of a six foot tall person is 24mm high, or just under one inch, rather large for a human pupil.

[maq1988]

I won't comment on eye science, but having read the piece my thoughts between the first and second picture is multi-parted. The first feels more real.

 

The first looks like it was taken in natural light, this provides a softness. Yes the scenic detail is there as well but basic, giving rise to weathered track and the corrugation of the building in the background. There's shadow, and greenery. The colour of the chute is okay - it's not wild pink or something untamed. Dark grey isn't amiss here. This adds to the realism I perceive.

 

The second photo, of 2296 is bland, I'd say unnatural light showing a harsh bare trackbed. The building behind is basic with no detail, no weathering. No shadows. The only colour is the stark red of the wagon, and the yellow of the locomotive. The eye is forced to notice these objects.

 

You mention the block background for photo 1, whilst it's extremely flat it is overpowered by the scene and colours in front so my eye isn't drawn to it. The second photo background is busy, multiple shapes and colours. The immersion is lost.

 

Perhaps a lack of depth of field is due to the first photo having tracks clearly at the front of the picture with the siding behind giving  feel of depth, and the second has little reference in the foreground to how far away the track is in the scene. 

 

There's also more '3D' in the first, the angle taken just above the chute gives an idea of size, where the only real thing to compare to in the second is maybe the roof, I assume it is meant to be pitched but it looks almost vertical.

 

EDIT: To those who can 'see' what the finished scene in picture 2 is meant to be have a better visual imagination than I. From that photo I would struggle to understand what I was meant to be seeing develop.

 

Anyway, I'm rambling as well and have about as much knowledge on the subject as a wet teaspoon.

Edited January 16 by maq1988

[Sotto]

This is a really interesting topic. My standard of what is 'acceptably realistic' is quite low by some standards, and for me it is easier to say what spoils the effect of realism in an otherwise  acceptably realistic model. I'd say the main one is inconsistency- the standard of realism needs to be similar for the whole landscape. 

Then there are scale errors, such as using a 1:50 scale truck on an O Gauge layout.

To get better than 'acceptably realistic' a layout has to take you somewhere. I remember seeing Pempoul at an exhibition; it just shouted rural France, and I wouldn't have minded if no trains ran. It is still the layout I have most enjoyed seeing.

More on topic, I'd agree that eyelevel is most realistic, but in an actual layout compromise is needed, otherwise foreground items obstruct the view.

[tom s]

Whenever a model railway photo is posted that tricks me at first glance into thinking it is a real life reference image, it is the lighting that does it. Getting the light temperature and diffusion right is probably an aspect that deserves more thorough review with the products available. 

Panel lining products that are common in military and automotive modelling, or washes in wargaming painting can also replicate the shadows that just don't naturally scale down.

Edited January 16 by tom s

[Sotto]

17 minutes ago, tom s said:

Whenever a model railway photo is posted that tricks me at first glance into thinking it is a real life reference image, it is the lighting that does it. Getting the light temperature and diffusion right is probably an aspect that deserves more thorough review with the products available. 

Panel lining products that are common in military and automotive modelling, or washes in wargaming painting can also replicate the shadows that just don't naturally scale down.

Panel lining can become an obsession though. I have a (waiting to be built) 1:350 U boat. I have watched a video by a military modeller which shows dramatic panel lining which to scale would mean panel gaps a foot wide. I have also seen photos of the real thing taken from a similar apparent distance- not surprisingly, no visible panel lines!

 

I do agree though that there are some astonishingly high standards amongst military modellers. 

Edited January 16 by Sotto

[MartinRS]

The prototype is often seen from some distance and not unusually from a vantage point higher than the railway. As far as realism goes I remember seeing the layout Gransmoor Castle which mimicked this view. Its high level of detail and rolling scenery worked well. Some years later I read in a computer magazine (or book) that a virtual reality effect started to trick the brain when a constructed object (in that case a computer screen) occupied a certain angle of the observers field of view. (IIRC it was 85°). I have since wondered if the effect could be replicated with a large (floor to ceiling?) layout. I have no intention of constructing such a layout though I suspect the impact on the brain would be interesting to experience.

[john new]

26 minutes ago, MartinRS said:

The prototype is often seen from some distance and not unusually from a vantage point higher than the railway. As far as realism goes I remember seeing the layout Gransmoor Castle which mimicked this view. Its high level of detail and rolling scenery worked well. Some years later I read in a computer magazine (or book) that a virtual reality effect started to trick the brain when a constructed object (in that case a computer screen) occupied a certain angle of the observers field of view. (IIRC it was 85°). I have since wondered if the effect could be replicated with a large (floor to ceiling?) layout. I have no intention of constructing such a layout though I suspect the impact on the brain would be interesting to experience.

Fully concur. I mentioned somewhere ages ago in a past thread one of the railway magazines I read in my youth, so late 50s or early 60s, covered viewing angles and perception.  If you imagine a triangle with the eye at its apex it widens as it leaves your head. IF you break that triangle with a close in but tall object such as a backscene board, low relief buildings or similar (i.e., tall & on a narrow baseboard) it looks as big as a flat wide baseboard with low level scenery. All comparable to that idea of the 85%.

 

Re floor to ceiling layouts wasn't that partially done IIRC with the Chee Tor layout in N and by Robin ?? De-Fraiesnet* in his large narrow gauge layout. The opposite was the MRC layout (Chiltern Green?) in that was very wide to get the same effect. What they have done too plus other effects with Copenhagen Fields.

 

* not sure how his name is spelt.

 

Edited January 16 by john new
Added an extra note

[The Johnster]

Eye science is objective about this stuff; I'm not. 

 

There are, I would suggest, two factors in our decisions as to whether a model is realistic or not, and I would not presume to suggest that the order in which I am going to mention them is indicative of their order of importance, it's just that I can't write the tow at the same time.  One is objective, and is scale, and I'm including detail and liveries in that.  If a model is built to exact scale with all the detail it must be correct and hence realistic, right?  Well, like Rogers & Hammerstein said, 'It ain' ness'rly so'.  It'll be realistic as a model, a single standalone object, but unless there's a setting, a layout scene, that frames and explains it's existence, that'll be as far as it can go.  And as soon as it is given a livery, subjectivity floods in like a dam bursting; pinning down exact colours is next to impossible, given that all methods or reproduction are imperfect, and that's before weathering, fading, weather conditions, atmospheric clarity or lack thereof, ambient lighting, the effect of layout room lighting, colour cast and intensity and all sorts of other factors, including individual human perception.

 

This leads to the second factor, which is subjective and holistic.  One is presenting a scene, imaginary or prototype, and it sort of doesn't matter as much as we'd all like to think if it is actually realistic, so long as it can convince us that it is realistic.  This is where disbelief suspension comes into play.  Lighting, sightlines, proscenia, scenic breaks, optical illusions, viewing height and position, ambient colour cast, all sorts of smoke and mirrors, in fact the greatest extent that we can possibly contrive to avoid realism while suggesting it by artifice.  If I can view a scene on a model railway and convince myself that it is real, but small, somewhere else, and at another time, then realism has been achieved, though reality hasn't. 

 

It's delusional of course, but I can do it on my layout despite the 00 compromise, curvature that would be unlikely on any main running line, tension-lock couplings, and all the usual compromises.  And this from a man who refuses to run anything directly out of the box, all must have at least a wash of weathering mix to take the new off because it's 'unrealistic'.  It works, it would work better if I could model better or I could afford DCC, but it works well enough and I'm improving it all the time.

 

Now, this sort of modelling philosophy can be a sloppery slip.  It requires as much discipline in it's own way as museum-quality scratchbuilding, in fact in one sense it requires more, because the scale-perfect modeller only has one task, to recrate reality as closely as possible.  The more holistic approach brings with it the yawning chasm of accepting lower standards than you have to, and convincing yourself that 1960s Hornby Dublo or 1980s Lima is fine and provides a realistic experience on a layout, and holding that modern RTR is overpriced, unneccessarily detailed, delicate, & complex, or a blatant display of conspicuous consumption or elitism.  It's not, but one needs to be rigourous in establishing the standards level that one can work within and holding to it for the best 'holistically realistic/realistically holistic' experience. 

 

Disbelief suspension is a tricksy thing, and, as with Rule 1, with great freedom comes great responsibility.  A layout is to an extent a work of art, or at least artifice, and to another extent the function of objective and subjective elements of creativity.

[DCB]

Interesting  thread.   I am long sighted, so I need to wear reading glasses for reading and  strong glasses for close work/modelling.   You can have the best  most detailed letting and fine detail on an exhibition layout and I'll never notice. but  a coach1mm too high at the buffers or cant rail, or a gap where there should be corridor connections, even excessive gap between coach ends jumps out at me and looks wrong.   Modellers struggle to get realistic lighting for a layout, overcast with 10/10 cloud cover might be doable, but anything  where the sun shines will provide shadows,
Shadows tell me which way full size trains are running.   Frequently layouts have a  lighting rig giving a representation of bright sunlight but without shadows.   
For realism the angle of view is  interesting, much like cropping photos I think the closer you get to the model  the better as the background blurs out,  The Gorre and Daphetid had floor to ceiling scenery,which disproves my theory,  as does my garden line as focus is on the moving train so the background blurs  out but  as I stop rambling I feel the talk of viewpoints  is irrelevant as there is no actual  fixed defined viewpoint  viewers will see the models from different viewpoints, unless it is through a lens, camera, web cam mirror etc.    Many of my favorite views of the layout are photographs from locations where the human eye cannot access...

[Nearholmer]

Multiple things going on at once.

 

IMO, the biggest factor is the willing suspension of disbelief, or put another way, deliberate self-delusion. 
 

When looking at a model railway in reality, we all know we’re looking at a model railway, whether it be The Vale Scene at Pendon, or Triang on pan-cake flat baseboard and with 13” radius curves. So, step one is to choose to believe. Go to Pendon, look at the Vale Scene, and look at it with a very cold eye, and you will find that for all the incredible science, artistry and years of hard work that have gone into it, there is no way it truly fools you ……. It doesn’t feel like floating in a hot air balloon over The Vale of the White Horse, it feels like looking at some small things. Conversely, crank-up the self-delusion, try to return to a child-like state where there is no boundary between objective reality and the imagination, and a yard-long train of plastic Triang whizzing round an oval can seem to be a majestic express thundering along a smooth mainline, and the carpet can actually be Shap Fell.

 

There is other stuff going on, for instance around having the model scene fill the visual field, so excluding real-world distractions, so if it’s a small scene, put a broad, dark-neutral border around it, and get viewers to stand close-up, which is the Rice-Cameo approach, inspired by theatre presentation (turn the house lights down and the stage lights up; get everyone’s full attention on the stage).

 

Then there are photographs ……. and I’m convinced that “different rules apply” from when looking at model scenes in reality, probably because the photographer, rather than the viewer, has firm charge of the field of view, and because there is less engagement, probably a different operation of self-delusion, and the things I mention below.

 

Eye, or more probably brain, science almost certainly factor into it somewhere. I guess that the brain calculates the size of objects using an algorithm a bit more sophisticated than the angle subtended at the eye (big = close; small = far away); it’s probably registering the changes in focus distance too, through muscle feedback. Scanning a 3D model scene will involve different alterations in focal length from scanning a real scene, and that may underly some of the difference between looking at a photo of a model (fixed focal length), and a teal, 3D model too (changes of focal length, but all at the ‘near’ end of the scale). Shadow angles from artificial lighting are another subconscious clue, as are absolute stillness, lack of reflections of the sky, dissonance between what the eye is being presented with and what the other senses are detecting (scene of frozen wilderness in model form vs. Hot, stuffy, noisy exhibition hall!).

 

It’s all in the mind, conscious and subconscious. The eye is only the light sensor, it’s all that software between the ears that’s interpreting and interpolating, and chucking out an output that says “real” or “model”, and the willing suspension of disbelief biases that output channel.

 

 

Edited January 17 by Nearholmer

[Nearholmer]

PS: I really like Johnster’s take on this, and would move it on a step by suggesting that its even “worse”, that the game isn’t even to suggest realism by artifice, that it is actually to evoke emotions, and that suggesting realism by artifice is but one way to do that; there are others. My “Deliberately Old-fashioned” layouts seem to do a surprisingly good job of evoking the 1950s; when people see them, they “feel” the 1950s, but they sure as heck aren’t realistic models of railways in the 1950s, even by artifice (of which there is very little aimed in that direction)., so most of it is about them looking a bit like model railways built in the 1950s.

 

I went to an RSC production of Les Liasons Dangereuses years back, and they way the setting managed to evoke all sorts of things was astonishing, but it was in no way “realistic”, it consisted of a number of white curtains hung at the back of the stage, and fantastically clever lighting. There was no image projection, just the use of lighting angle, colour temperature, intensity etc, yet it managed to convincingly convey things like “indoors on a spring morning, with windows open and a fresh breeze coming in; optimism and opportunity”, and “candlelit interior of chateau, late in the evening; claustrophobia and unspoken threat”.

 

 

Edited January 17 by Nearholmer

[The Johnster]

A good example is the scenic break into my fiddle yard.  There isn't one as such; it's an awkward location, and what happens is that the scenery just sort of gives up, like Bonnie Prince Charlie at Derby.  The lighting is arranged to distract one's attention away from the area, and while I fully intend to do something about it one day, past attempts have failed to cut the mustard and I now just sort of mentally tune it out, so it actually works very effectively...

 

A lot of the space on real railways is mostly empty most of the time, and our layouts are overcrowded.  The throat end of the colliery yard on Cwmdimbath, for example, is somewhat bleak and devoid of visual interest, for which there is actually not much room and little prototypical need beyond perhaps a shunter carrying a pole while he walks to change points.  One would normally address this with cameos, but it would be a pretty bleak area in real life so anything much on the ground will look wrong and be overkill.  I'm chucking ideas around in my head involving maybe a pipe bridge or something to lift the flatness, we'll see.