Raptor 3D Project - Research and 3D Setup

I believe that no model is better than the research and documentation you have in hand. So for me, there's always a lot of "immersion" at the beginning of a compex project. I do a huge amount of research, and spend a lot of time reviewing that research, before doing anything in 3D. This gives me an overall feel for the aircraft, and some time to think through the problems I'll have to solve. Go carefully at this stage, and you may be able to avoid mistakes that could demand corrections later. So, a few words on that is appropriate here.

Several years ago, I did an article for my archive site, (http://www.nextcraft.com/plastic_to_rc01.html) showing modelers how to slice up a plastic display model, to come up with cross-sections, probably the most sought-after bit of documentation we'd all like to have. Over the last 12 years, I've also written to Lockheed-Martin, and several offices associated with the Air Force. Each of them so far continues to basically say (nicely) "No way". So, a little creative thinking is in order, and some new methods had to be invented.

A unique "collaboration" of sorts...

I built every 3D model you'll see on my sites, entirely from scratch, doing my own research and modeling. For the first time, I went looking for a little help, in the form of an existing 3D model I could modify. As a 3D modeler, I'm aware of many 3D forums, galleries, and online "stores" that sell 3D models, so I looked at all of the 3D models of the F-22 I could find. Doing that not only gives you information about the "competition", but points out various methods that previous modelers have used to create this aircraft in 3D. You always learn something along the way.

Thanks to immersing myself in the documentation, I can tell now, using my eyes alone, when a Raptor model is "way off" or "pretty close". After reviewing those models, and even obtaining some of them, I was frustrated. It seemed that most were inaccurate enough that I might as well start from scratch. Then, I found a truly excellent one, by someone I'll refer to here as "the original author". (He prefers to be anonymous at this time.) I described the project to him, and obtained his permission to use his model as my starting point. I also obtained his permission to use it in this tutorial, and even to eventually sell it, when the model is more than 70 percent rebuilt. That's quite fair, and is a very rare attitude today, and I cannot thank him enough!

The original model was created in "LightWave", and sent to me in both it's ".3DS" and ".DXF" form. At the time I started the Raptor project, I plannd to upgrade to LightWave, but in the meantime, am still using "Carrara Studio". So I imported both file types, selected the .3DS version, and converted it into what Carrara calls "Vertex Models". (polygons) The image below shows part of the original author's model in Carrara, with his texture maps removed. It's quite elegant, thanks to the nicely-organized polygonal structure. I can't live with myself if I build a crappy model and disguise it with a texture map, so good, solid, polygonal structure is important to me. I tend to produce high polygon count models, relying on actual geometry, as opposed to textures and shaders.

If you understand the images here, you'll see why I was so grateful. Carrara is one of those 3D programs that will sometimes "bend" rows of polygons. LightWave apparently keeps this issue squared away, and makes editing a LOT easier.

First step: Disassembling the model and regrouping it into separate sections

As a 3D modeler, you can group and export your finished model any way you like. You can group the entire model into a single object if you like, or you can export it still grouped into logical subassemblies, like "landing gear", "wings", etc.. In this case, the original author's .3DS file only had about 6 groups, sometimes with quite a few parts grouped together, as in this image.

For me, these parts had to be regrouped into "canopy", "canopy frame", inlet left". inlet right", and so on. Regrouping and re-coloring the model to this stage took several hours. I then made a duplicate of the file, and saved it, in case I need to come back to it. All work begins with a COPY of the original file.

I'm using an overall color (below) that's similar to Lockheed-Martin's factory "primer" so that the model in progress can be compared to factory photos. Later, I'll add "serious" texture maps and other coloring, but for now, I'm concerned with the external geometry, and simpler colors help.

How to achieve "scale accuracy" on a plane with little available documentation:

First, I collected every shred of information I had, in the form of isometric drawings. Naturally, you have to be careful here, because, due to security concerns, some things are omitted, or even changed, in some of these drawings. So, the first "proving" step was to scale the various items I had to the same size, and then use PhotoShop to stack them up in layers. (images on next page) That way, I could compare differences between the drawings. While you're doing that, you can also compare the overall measurements to the ones published for the public. The most widely circulated specifications for the F-22 state that the wingspan is 44.5 feet, the length is 62 feet, (or "62 feet, 1 inch") and the height as 17 feet. I take that as true.

There's a particular thing you have to be careful with, when you start to model an aircraft, based on drawings. It may sound silly, but trust me... If you don't make a decision on this early in the process, you may develop a large headache. It's called the "squat angle", and simply is the angle that the airplane sits at while on the ground. Take a close look at the side view and front view drawings. Do they show the landing gear oleos compressed, and/or do they show the tires "squished down" a bit from the weight of the aircraft? It may cause the aircraft to sit at 1 degree nose-down, for example. If you don't notice that when you start, you may find yourself building every 3-dimensional part with an error equal to the "squat angle". I believe it's better, in the case of aircraft, to find the "aerodynamic zero" cruise angle for the plane, draw a line through that, and align the model's pitch to that "zero" angle for the whole modeling process. At the very end, it's a simple matter to tilt the whole model to the correct ground angle. Simply put... "Build at zero degrees... Tilt later."

Aircraft, ships, cars, and other "vehicles" often have something in common, which is bilateral symmetry... meaning that (aside from details we'll do at the end) they are mirror images of each other. That alone can save you HALF the work at times, by allowing you to work on just one half of the model, and duplicate it later. That's a good thing.

Did aliens build the pyramids?

One of the many misnomers that conspiracy-oriented groups like to exude is that the ancient Egyptians couldn't have built the pyramids, because they're "too precise". Did that job require anti-gravity machines and lasers? Some would actually answer "Yes", but that's an uninformed point of view. Consider a few of these more rational approaches instead:

• To make any surface smooth, sand it. You don't have to be able to "know", or state verbally that it's got an error of "three thousandths of an inch". You simply need to "feel the bump" with your hand, (and/or see it) and sand it more. An average, untrained person can feel tiny errors to a greater degree than you might believe, and of course, with training and experience, that improves.
• To find a perfect vertical angle, hang a rock from a string. To find a perfectly horizontal line, put water in a bucket. The surface is level. To create a straight line between two points, tie a tight string between them. (No aliens, government coverups, or lasers required)
• To draw a perfect circle, tie a piece of string to a pin, stick it in a piece of paper, place your pencil in the other end of the string, and draw. Simple.
• If you have photos of a thing, they will include shadows and reflections. So, in 3D, "paint" your 3D model similarly, light it similarly, and view it from the same perspective. You can then use the reflections and shadows on your 3D model to judge it's accuracy. Light travels in nice straight lines, so this is a valid method.
• Most of us can instantly recognize (in a movie, for example) when something is synthetic. Use that instinct. If you're really paying attention to your photos and other documentation, you'll intuitively know that your model needs adjustments. Trust your eyes, and make small adjustments, until your brain says, "That's it!"
• Understand that the people who really know the inside details on a project like the F-22 are NOT going to tell you those details. They not only will withhold information in general, but may actually tell you something that they know is wrong. It has to do with national security, so take publicly-stated "facts" with a grain of salt, and do your own research.
• One of my RC modeling heroes, Dave Platt, is known for building beautiful scale aircraft, and for his incredible hand detailing on the markings and paint. When asked about the various shortcuts used to do these things, he said "I've tried all those things, but if you really want great results, then you might as well just start in and do the job". In other words, If you want to do something difficult, do it. Don't be overwhelmed by ANY level of detail or complexity. The most complex task can be broken down into smaller, simpler steps.
• Simply put... "Use the force, Luke". Your eyes, connected to your brain, can recognize a huge amount of information from photographs and drawings, by simply studying them with intensity.

The rebuilding process starts on the next page. ("Edits 01" link, below)