I wanted to make an 'architectural' miniature of the vessel, like the industrial models Nakatomi display in their plaza. This makes sense, it doesn't have to appear realistic or weathered, just detailed and accurate. It also lets me work at a scale thats too small for a useful studio model.

Instead of kit bashing the usual way with styrene parts, I started looking for random things from hardware stores and supermarkets. I had assembled a bunch of pieces that would go well with each other.

And here they are all laid out in raw form, including the polyurethane party poppers, which are a real pain to work with

But I discovered a big problem with this approach - consumer plastics.

A lot of consumer and hardware plastics you come across are made from polyethylene or polyurethane. The one thing these plastics have going for them, is that they are chemically unreactive. Thats great for storing bleach, but it means that its quite hard to paint them; the solvents in the paint will not seep into the surface, and the pigment just flakes off when dry. The same thing happens with glue, it just hardens and cracks off.

Primer just chips off at the slightest brush

I tried a whole bunch of glues, even the all-plastics one with the special accelerator which is supposed to soften up the plastic and glue the ungluable, it didn't work. It finally struck me when I looked at the bottle of glue itself, and realised the container it came in was the very material I was trying to bond. So, back to the drawing board.

It was also dawning on me that the particular design I had for this ship was going to require massive duplication of parts. If I wasn't smart about it, it would actually be duplication of labour. I started looking at alternatives. One idea was making masters and casting the parts, but that brought its own problems. So I started looking at 3d printing instead.

Up to this point, most of the 3d prints I'd seen used the powder sintering method. This results in a rough pourous material, definitely not what you want for a small scale miniature. But Shapeways were also offering a material they call Frosted Detail. This process uses a laser cured acrylic polymer, and is much more suitable for miniature work.

So I set about designing the parts I needed in blender. Because of the high cost of the printed parts, and the volume I would need, I was still going to need to do the 'bulk' parts another way, and just use the printed parts as greebles.

My modelling skills are very rusty, but this is the kind of thing you can just make up as you go along. I measured out the real world piece as accurately as I could, and got to it. There was a lot of back and forth, getting the mesh printable over several failed submissions. But this was the final result:

The parts were modelled 'in place' on a stand in for the real world piecesThe parts need a specific layout for print, with the right amount of separation and orientationI cant complain about the packaging from Shapeways, that huge box in the background housed these tiny pieces in all sorts of bubble wrapThe parts have been cleaned and dusted with primer

The parts have a release/support material on them, similar to wax, which must be removed. The best way to do it is with an ultrasonic cleaning bath, the kind used for jewellery cleaning. I just washed them with soapy water a few times and let them air dry.

The accuracy of the parts was impressive, they had a tight friction fit in the low grade, oddly shaped party popper.

I had also arrived at a possible solution for the polyurethane problem. It was hacky, but it worked. If I wrapped the parts in PVC electrical tape, I could glue and paint to the tape surface, and it would stick to the underlying part by tack alone. I could also use tiny pieces of tape to add detail.

Ironically, when I was typing 'hacky' in the paragraph above, it got corrected to 'tacky'The parts come together quite well, and hide the fact that its a party popper stuck to a thread spoolWith a bit of mood lighting for good measure

Putting it all together into a test fit, it was starting to take shape. The next step was working out the internal lighting, and that was going to make all of this seem like a walk in the park.

Those surface mount LEDs look mighty small..

One of the things that I'm really looking forward to doing when I make the Iapetus movie, is doing practical miniature effects for the exterior spaceship shots. One of the earliest things I designed (back when this was an animated film) was the ship..

iapetus_98_1iapetus_98_2

It's intended to be a colony ship, holding about three thousand people in the five cylindrical habitats, and the engine is a long linear array of accelerators. In real life, its about 3km long.

When creating convincing miniature effects, the real trick - apart from creating a highly detailed model - is miniature photography. The key to the effect is focus. In order to make a small object seem large, all of it must be in focus at the same time. You can see this very clearly when you look at the reverse of this principle. To make a large object seem very small, you blur the foreground and background, an effect known as tilt-shifting or miniature faking. This is the look you want to avoid at all costs.

A real train made to look like a miniature, the goal of miniature photography is to do the reverse (photo by Scot Campbell)

The solution is to make the aperture on the camera as small as it will go. This reduces the light that reaches the sensor, so you need long exposures. But, long exposures mean lighting effects (like windows, engines) get over exposed, so you have to do those in separate passes. That's where motion control cameras come in. They can perform the same movement as many times as needed to capture all the elements.

For the shots I wanted though, all I need is a simple pan. As it happens, there are motorized sliders out there which will do the job. They are usually used for time-lapse photography, but the principle is the same – capturing multiple long exposure images over time. It turns out the guys who made C had the same idea, so we now know it works.

There are only a few simple shots required for the trailer

Ok so miniature photography is important, but you still need a miniature to photograph. Because the design of the ship contains a few structures duplicated many times, I can take advantage of that and only build one of each. These objects can be physically repositioned, and a pass can be photographed on the motion control slider. When you composite them all together in the correct Z order, they appear to all exist at the same time.

The same object is photographed at different positions

To see if this effect would work, I decided to build a mock-up of one of the engine modules out of paper. I also put some lights in there to test how the engines would appear on screen. After photographing each element separately, I composited them together against a background. The result proved that the effect was convincing, the objects all appear to be part of a larger whole.

A paper model was constructed for the camera testSome internal lighting was added to see how it would work when photographedThis multiple exposure shows the how the principle works

The result works pretty well, it appears as a single solid objectA view from the rear to see what the engine lighting looks like in context

So I have a way to do the large scale close up miniature shots, but for distant establishing shots of the ship, a smaller model will be needed, large miniatures don't look 'small' enough when shot from a distance. I was originally planning on just constructing this from scratch using bits and pieces, but then I discovered 3D printing over at Shapeways. I created a quick mock-up of an engine pod in Blender and printed it out.

Thrown together quickly in BlenderAnd a week later it exists in the real worldSize comparison with the larger papaer model

This is printed using the coarse grained 'strong and flexible' material, a kind of nylon. Because there is no edge smoothing, and the mesh doesn't have enough faces, the curved surfaces are stepped. The Ultra Fine material with a high resolution mesh and a good paint job would be more than adequate for a long distance shot.

And that takes us up to today. I'm continuing to build out the small scale model myself, but I'm going to get a professional model maker to work on the larger scale piece over the next year or two.