Scene Composition: A Case Study

Scene composition is always an interesting game to play. Yes, it is a game; a hobby by itself if not an art form that can be pursued, enjoyed and perfected. However, we often think about scene composition from a landscape and structure standpoint, forgetting for a moment trains are also part of the setting and must be taken in account. Nobody would shoot a movie or draw a comics without taking in account how to place the actors.

I recently commented how I prefer to be immersed in a small scene where operations can be enjoyed from a single point of view. The idea is basically to enjoy the action as would happen in real life and appreciate how large steel cars and locomotives are passing by. Over the year I’ve built many mock up layouts and found out optimal vision angle in such situation works best with a layout scene about 3 to 5 feet long. Over that, you need to move around too much and it starts to break the suspension of disbelief. I suspect this will vary greatly from individual to individual, but let’s call it a good sweet spot in my case. A few of you must recall my old Quebec South Shore Railway layout was based on a similar concept and, interestingly enough, lost its charm the moment I tried to extend it, thinning down the overall immersion effect.

When that parameter is known, which is the size of your "frame", scene composition can start. By scene composition I mean the structures and landscape will shape a visually compelling scene that also helps to support the illusion a larger world. Many tricks can be used including vegetation, overpass, foreground buildings, backdrops, etc… But one’s must keep in mind the action has also to be taken in account. If your wonderfully detailed and weathered locomotive is half the time hidden in staging or somewhere else, maybe you won’t get the fun you wished for. It must be noted this becomes quite relative when dealing with larger layouts and trains, just has it happens in real life. For this reason, I’d like to stress I’m discussing smaller switching layouts here and not larger ones which work under fundamentally different sets of parameters.

Boston and Maine Station, Standstead, QC (postcard)

With that in mind, we can now start to develop a concept that focus on a small scene will making sure most of the action is done on the layout visible portion. For this example, I revived a layout concept I designed a few years ago. It is based on a small Boston & Maine Railroad branch line in Stanstead, QC that operated from the 1900s to the 1930s (B&M pulled out of Quebec in 1926 and Quebec Central ran the line until its demise). The branch line was about 3 miles long and was used to link Stanstead and Rock Island to Massawippi River Railway main line connecting B&M to Lennoxville and Sherbrooke. Both locales were then bustling towns which fueled their prosperity by extensively quarrying and exporting granite both for the domestic and american markets.

As suspected, trains were extremely short but it also meant the small terminal in Stanstead would be quite interesting. I only found a few photographs of the place online, but most of them are quite interesting and packed with details. Business seems to have been quite good around the railway and car loads were diverse. While my goal isn’t to reproduce faithfully the place but only to use it as a factual starting point, I'll took a few liberties when designing a track plan. That said, you can easily imagine a passing track, a mainline which hits a bumper where some cars can be spotted, a small standard passenger depot, a freight house and a team track. A few hundred feet away there is a 52'6" Sellers cast iron turntable suitable for diminutive 4-4-0s and a enclosed water tank. Quite typical, quite mundane yet full of character.

Since the scene is to be about 5 feet long, it becomes quite clear the locomotive facilities (turntable and water tank) can't be include on the main module but rather integrated in the fiddling area. What we should care about is the end of the line, where most public-related structures stand. It means we have less elements to play with, keeping the selective compression manageable and structure density on par with the prototype. It must also be noted Stanstead station used to be located on a small topographic depression, which gives us the chance to use landforms to frame our subject.

After a second thought, turnouts should be at least #7.

The scene is composed by setting the largest building – the depot – on the background. First, because it is convenient and a good place to show off a good model, but also because a locomotive running in front of a station is always a sure win in term of visual impact and a good place to shoot interesting pictures. I love when station are in the foreground, but it isn't a very practical approach when heavy switching occurs in an already cramped space.

Near the end of steel, I place the freight house on the foreground side of the main line. Once again, many reasons are at work. First, it provides a good way to conceal how the track meets the backdrop. Due to shortage of space, I prefer to hide it in a convenient way such as this one. It also provides visual interest because you can see a car lurking behind the freight house, giving the impression stuff happens beyond our zone of interest. In term of composition, it also balances the station in such a way you get buildings on both side of the track, creating an interesting depth in what could be a flat scene. Finally, having the freight shed on the front brings life to the layout because the structure can interact with the outside world. It is expected to find vehicles and peoples around there that comes in contact with the railway. It is both visually more interesting and provides more photographic opportunities.

On the other side of the layout, where trains enter the scene, adjacent lands rise to create a visual bottleneck between the layout and the fiddle yard. This is an old and efficient trick which is often done with an overpass. Other tricks would involve a structure on the foreground or thick vegetation. All of them are equally interesting and it’s a matter of choosing something that makes the scene believable. In the case of Stanstead, an overpass would be completely out of place and it's why I think simply raising the adjacent terrain and adding proper vegetation will work better while keeping the impression the outside world isn't far away.

Framing train operations from a visual perspective

Finally, fine tuning must be done. When I first drafted the idea, I created a typical consist made of two freight cars and a passenger cars. Given locomotives would be small steamers such as 4-4-0, 2-6-0 and 4-6-0, a typical train would be about 28” long at best. Knowing that, we now need to create a composition in which most actions performed by the locomotive happen on the layout. As much as possible, time in the fiddle yard must be minimized. A key element is where the train is located when arriving and departing the station. Basically, it means any train serving the passenger depot must be fully visible in the scene. Having a departing train locomotive hidden in staging would be a no go for me. Thus by tweaking the depot location and the platform, we can find a specific spot where both trains are perfectly displayed on the layout. No magic or trick here, it’s just a matter of working with real dimensions and understanding how trains will move in and out. In that regard, I mainly based my approach by mimicking Trevor Marshall’s excellent Port Rowan layout in S scale.

As you can see, the final result is a compact yet realistic layout with a simple but prototypical and functional track plan. By using a turntable in the fiddling yard, a lot of space can be saved to keep things breathing on the main module. To some extent, if returning locomotive isn’t required, a simple sector plate could do a terrific job. Given the dimension, such a layout could easily fit an office room or a shelving system. Care and attention to the finishing would make it blend seamlessly with its surrounding and could truly be considered as some piece of art. The fiddling yard could even be hidden behind a cabinet door when not in use if it isn't scenicked. In the case of an exhibition layout, the fiddling yard could be enlarged and enhanced to stage multiple trains.

Many themes can be explored and such small layouts have always been great ways to reach a certain level of proficiency. British railway modellers and others around the world have proved us this kind of formula – when it fits your needs and interests – are a joy to build and operate. As I often say, I will never advocate a single type of layout. However, I think it is interesting to see how small layouts can easily be turned into piece of arts and great exercises in scene composition. They make your brain works and push your design skills to the limit, which is always a good thing. This will come handy when designing larger layouts. Mastering the basics is always easier when working on something small rather than spoiling resources and building up frustration when experimenting with large projects.

As for me, if I had to build such a layout, I would go with an earlier era just like I decided to do with my Temiscouata project. The small rolling stock is perfect in term of space and small steamers perform well with short trains (just like the prototype). Many interesting cars are available on the market and old craftman kits can be purchased for a nominal price and improved greatly using modern details and techniques. Labelle passenger cars are perfect for that era and can be customized to fit a particular prototype. Accurail is also producing excellent plastic injection kits of classic 36’ boxcars which are a joy to kitbash and details. Finally, small steam locomotives made by Bachmann, Roundhouse or even the old IHC can become great project and generally perform well with some care. Older brass steamers in need of restoration can also be found for a good price. Given the layout small size, all these projects are manageable because you don’t need a ton of rolling stock to support operation.

Improving Walthers CN Difco Color Scheme

Strong from my previous experiment with color filters, I decided to see if I could improve an old late 90s Walthers CN Difco Dump Car. I must stress this car isn't prototypical even if it is quite close to a few cars CN owned. Most CN dump cars had completely different panel designs, larger cylinder and many of them were noticeably shorter. To add to the injury, the lettering has nothing to do with what was common on CN. Back then, Walthers used a very dark burgundy color instead of boxcar red for CN cars, which gave them a weird appearance. Fortunately, new releases have a more typical shade of brown and far more prototypical lettering.

Parts sprayed with a new shade of brown and lettering cleaned

In the past, I've weathered this car twice. It explains why the side panels are full of dents even if it doesn't make that much sense. But back when I was in high school, it was the best I could do. I also recall having painted small graffiti has seen on a car stored near my school.

Anyway, for the sake of this experiment, I decided to keep the wrong lettering and accept the car was a stand in. However, it was clear in my mind the wrong color - which is a major detriment - had to be addressed. To change the color, I sprayed the model with acrylic paint with a correct shade of brown. To get it right, I didn't rely on official colors, but rather mixed my own color following prototype pictures. It was evident these cars had badly faded paint and I wanted to replicate that particular look.

When the paint was dry enough to touch the model, I used a stiff paint brushed soaked in isopropyl alcohol and cleaned the lettering. It didn't take long to see the white shines again. From there, I used various shades of highly diluted oil paint to modulate the brown color. Raised parts were highlighter with light and warmer tones to make them pop and give the car so depth. The same was done on the air cylinders.

The car interior was weathered after I carefully studied Difco car videos on YouTube. Several people have posted online videos showing cars in action. When the load is completely dumped, it is easy to see the weathering patterns and how the steel is polished by the sliding loads.

When dry, everything was clear coated and standard weathering techniques were used to dust the model. Finally, a last clear dull coat was applied but this time, I used a custom mix of 1 part Tamiya flat base and 3 parts Future acrylic floor finish. I like this mix because it create a chalked effect on the model that suits perfectly well fade paint.

At the end of the day, the model discrepancies aren't gone, but at least, it visually blends better with the rest of the fleet. Given this model is rarely used in operation, I'm not that much concerned about prototypicalness. If I ever want a better CN dump car, I'll simply scratchbuild one.

To Each His Own Point of View

The way we interact with layout and model trains varies a lot among modellers. Like any subject matter on Earth, many people will like the same thing but for different reasons.

This is how I like to see my trains and be a part of the scene.

Last Friday I had the chance to operate the layout for the first time in about two months. A simple train from D'Estimauville to Clermont, switching en route the team track and the large paper mill.

While my freight manifest crossed large expanse of territory, I really was sucked in the miniature world when I was sitting in front of the 3 feet long team track and spotting 3 bulkhead flatcars full of lumber and a loaded ballast car.

Over the year, this feeling occurred many, many times: I had most fun while watching my train at eye level on a relatively small scenes (about 3 to 5 feet maximum). I can recall doing the same with my childhood train layouts. Regular readers knows too well how I have a fascination for smaller layouts and I must admit it isn't just for the sake of pragmatism. I truly need very little to be captivated in term of model railroading.

However, today a weird idea came ot my mind. I generally push forward the idea that I prefer when a layout emphasize the linear nature of railroads, thus implying a good layout would be long scenes with very little compression. This, honestly, can't always be done and I'm not sure it yields as much satisfaction one would want due to many other constraint. However, I thought what if someone would build a medium-sized layout by creating several vignettes of the road, each large or small enough to focus attention on what matters. It could be likened to a series of cameo layouts. Since each scenes would be self-contain, the distance between each of them would no longer matter nor the way the blend together (which is always tricky).

I'm well aware this design idea is absolutely not new and many have implemented it over the past with good success. However, it has some merit: first you have the chance to model various small scenes with a focus equivalent to a good quality diorama. Scene composition and operation are what will determine the size of a specific scene. Also, given railroads are quite large infrastructures, I see no harm in cutting them into small vignettes since it is how we perceive them when we railfan them.

Another interesting aspect is that some scenes can be completed and fully put into operation while others can be done later when inspiration, time and resources are available. Basically, the layout grows as required and few small  good scenes can provide a lot of satisfaction.

One of the reason why I put forward this idea s because I always struggled with Hedley-Junction when it was time to blend together several scenes. At best I got mixed results, at worst I was conflicted over the end results. Also, since our main goal is to replicate operation on this subdivision, a lot of compromises had to be reached to somewhat replicate the real thing. Unfortunately, we ended up sacrificing many interesting customers that define the nature of the line or geological features that set the locale.

Don't be fooled, this ain't a vary on the "I want it all" syndrome, but rather the fact I often found out many scenes on Murray Bay subdivision could be modeled as small switching layouts but didn't work when implemented on our main layout. I could distinguish both approach as the traditional one that tries to replicate a miniature open space world or one that that focus on specific enclosed scenes. You will notice I stressed "enclosed scenes" because I believe just putting together individual scenes together in an open air fashion is tricky at best.

Also, in a design, the layout would become something that would be quite similar to a storyboard. Each scenes being relevant in telling a story about how the line operate. In our case, I can easily imagine the introduction scenes at Maizerets, a second scene in Villeneuve focussed on the cement plant, then a third one about Dominion Textile and Montmorency Falls, and so on.

Very small scenes like Léo Cauchon sawmill in Château-Richer, Ste-Anne's feed mill or Cap-à-l'Aigle pier scenes could easily be implement since most of them don't need more than 3 to 5 feets to be modelled with success.

Thinking this way also enable us to figure out many long stretch of tracks are unrequired if your intention is about focussing on a few scenes. Indeed, if the layout goal is different and based on "running" trains, particularly very large one, the answer will be different to fit best.

CN Ballast Hopper Completed & Military Models Weathering

I'm glad to report the CN orange ballast fleet is now done once for all and ready to enter service (when the paint will have hardened!).

The last step was about weathering the interior. To be honest, I didn't go very creative and I simply followed some basic instructions outlined by Ted Kocyla from the Waterloo Region Model Railway Club. This article is interesting because it is dedicated to ballast hopper and how crushed stone create a different weathering pattern than coal which is acidic and causes much more corrosion. Basically, the interior was airspray with dull aluminium paint, then followed a coat of dark dirt paint on the higher parts where less abrasion occurs. Some PanPastel dusting followed to replicate some fresh rust and dirt. The model was then dullcoated to seal the effect. When dry, I splattered thinned down Burn Sienne oil paint to recreate the rust pitting typical of exposed metal. A last coat of Future and Tamiya flat base (about 3:1 ratio) was applied to seal and blend together all the effects.

I've also been asked by fellow modeller George Dutka about more details related to the use of filters to modulate a base coat. This technique is widespread in military modelling, but rarely used on model railroads. Many of our techniques are somewhat similars, but I feel they are done in a more whimsical way than military modellers do and ofter, we make it much more complex than it should be.

As their name implies, filters are very light coats of color that are applied to a base coat to create slight variations of color. The goal is to tint the original color to create more depth and bring life to the model. It must be noted military modellers seem to categorize independently a lot of techniques we would put together under the generic word "wash". But I think they are quite right. A wash would completely flood a model to get an even coverage of all parts. This is what we do when we apply an oil paint wash over a model to get a grungy look. While very useful, this technique can get quite repetitive and generally fail to induce more subtle color variations.

Many of you have probably seen many railway modellers highlighting freight car panels by locally modifying each of them. It can be used to replicate decoloration, warping, buckling or highlights/shadows. Some use powders to do that, which can be a real pain when dodging the lettering. While a legitimate method, it seems to be quite inefficient. I did it once and called it a day.

With filters you can do that in an easier way. Your goal is to create a very light tint by diluting a little bit of paint. The ratio can be as low as 1:20 as you will wish to build up the color . It can be done with acrylic or oil paints. Oil paints enable you to apply the filter with a brush and gives you time to work the effect a little bit before it dries. Surprisingly enough, since a filter is extremely thinned down, colors such as pure yellow, orange, magenta, blue or violet can be used and create subtle but impressive results. According to prototype references, you can uses various filters on different parts of the model to make details pop or create peculiar fading patterns. Also, ready made filters can be purchased on the shelves. They have the advantage to be consistent and many major manufacturers offer them.

Toronto IPMS has some helpful videos on YouTube about filters and other useful techniques. I must admit I can't thank Mike Cougill enough for pointing out this series of videos. This one covers what I did with my hoppers. They have many online clinics which are quite useful and I strongly encourage you to watch a few. Renowned military modellers Michael Rinaldi also gives some great advices in using oil paints intelligently so it makes your life easier and your model quicker to dry.

This rust pitting effect would have probably looked better if done before decalling

Has you can see, theses techniques are hardly ground breaking. But as Michael Rinaldi points out, it's not about "new techniques" but rather about how you can drastically improve the results by refining the way you do it. To often model railroaders weather with a heavy hand and rely far too much on pure luck instead of controlling what they do. I believe many of our methods are OK, but poorly understood and applied as a "miracle recipe". Some thoughts should be put in what we are doing and I'm seriously starting to believe we should start to shuffle some building steps, including the moment we letter our cars. As I previously said, we are making our lives much more complicated than it should be.

On a funny note, when I was talking about that with a friend who models wargaming tabletop figures, he pointed out that model railroading rely a lot on prepainted model, which he surmises made us forget to break down the assembly and painting process. To be honest, I think he has a good point. We indeed try to make our models look like factory built RTR stuff before thinking about weathering.

And by the way, don't get me wrong. I am fully aware many railway modellers have created fantastic weathering methods over the year and share their knowledge on various forums and social medias. However, as great their work is, I find it disturbing to see this knowledge barely permeate in the mainstream and is branded as stuff for "geniuses" and talented people when it has absolutely nothing to do with it. Wouldn't a long series of advanced articles about that kind of stuff would make the hobby much more inspiring than covering another generic basement filling Plywood Central that doesn't fit most people's life? Given you only need to sit at a table with minimum supplies to create masterpiece, I certainly wonder which subject would have a better impact of modellers self-esteem and feeling of actually achieving something meaningful instead of being crushed by pipe dreams.

Some Details About Weathering Ballast Hoppers

CN orange ballast hoppers develop over the year a peculiar weathering pattern. Most pictures from various eras show them with a distinctive occurrence of rust pitting and abrasion on panels, keeping the ribs rather pristine. No streaking seems to play a major part in the weathering process but only a typical accumulation of dark dirt on the ends and under body. My I don’t know the real process behind the rust pitting, I wouldn’t be surprised it is the result of loose ballast stone dropped against the panel when the cars are loaded. To further my point, you will also remark the top horizontal stiffener is also quite rusted in the same central area while in good shape at the car ends.

Top: unweathered, bottom: with filters and weathering (yellow hues can be seen)

To replicate this neat effect, I decided to first fade the orange paint using a yellowish-tinted filter. Filters are used to alter the base color. While model railroaders have a tendency to completely fade a model, military guys generally do it by panels in a more controlled fashion. Different tints can be used to add variation and create steel panel buckling (modeller Tom Johnson is well-known for using this method to create hyper-realistic grain hopper weathering). In the past, when I weathered the initial fleet of Harlem Station cars, it is a method I used intuitively and which I thought brought far better results without having to flood models with heavy washes.

Light weathering and yellow + burgundy filters

Once the color is altered, I came back with the airbrush and added a very subtle coat of thinned rust paint where the pitting effect was the most obvious. From various pictures, it is possible to see rust leaching out of the pits, creating a muted coat of rust powder over the intact paint. When it was dry, I splattered thinned burnt sienna oil paint over the rusted spot until I got an effect close to the prototype.  Once done and dry, which didn’t take too long since I used very little solvent , regular weathering techniques were used to pop details here and there or add some dirt.

The interior is yet to be painted and weathered independantly

I previously said I wasn’t interested in replacing the grabirons and that I would take care of their appearance during the weathering process. To make them appear thinner and more prototypical, I weathered them heavily until they were almost dark brown. When dry, I used a brush loaded with Tamiya airbrush cleaner (or any other strong solvent) and cleaned the grabiron face to bring back the color. The result was a thin orange line while the thick plastic section remained quite dark, making it disappears.  It certainly doesn’t beat metal grabirons, but it definitely helps to conceal oversized molded on details. Also, if you look at prototypical pictures of CN ballast hoppers, it clearly appears the grabirons keep their bright paint in stark contrast with the dirty surrounding steel components.

Finally, a protecting matte coat was applied. This time again, I didn't follow the normal recipe but added a light tan tint to the mix before airbrushing. This helped to blend all the weathering effects together while brightening the overall model as if it was under the sunshine. It is well known we have to take in account that color does indeed scale and this is a good moment to do it with this last control coat.