When you’re making machinima, you have a bunch of virtual actors standing around in a virtual world doing, in the immortal words of the ILL Clan, virtually nothing. You have to inveigle^[13] them into doing what you want. You have two choices: control them with machines or with people.

Your first option is to use the game’s own artificial intelligence to control your actors, usually through some form of scripting language. Yes, you too can now control an army of robotic minions. Theoretically, this means you can make an entire film by yourself. The creator of Nehara made an entire three-hour movie this way in six months. In addition, your actors will work as long and hard as you will, and you have total control over their every movement.

The other option is to have your actors controlled by actors, or, at least, humans. Grab the mouse, configure WASD and a bunch of emote keyboard commands, and you’re off. It’s fast, it’s fun, and it’s very flexible: humans don’t give you syntax errors and they don’t crash, except possibly after a 16-hour session fuelled entirely by Jolt Cola and goodwill.

Next, you have to choose whether you will use in-game content (maps, characters, special effects, etc.) or new work.

In-game content is convenient, it doesn’t demand any artistic skill, and it’s normally of high quality. However, if you want to use all in-game content, you need to tailor your film to the content available.

A lot of groups end up creating their own content because they feel that the existing in-game maps and models don’t suit their movie. This is a very good idea if you want to differentiate your film, or if you want to make something totally different from the norm. See films such as Hardly Workin', Eschaton, or In The Waiting Line. However, don’t underestimate the amount of time this will take. Two of the films I mention took more than a year to produce, and the other had a commercial budget.

Of course, this relates to another choice: the game you will use. With the wide range of 3D games now available, most of them including editing capability, you can usually find something that will suit the story you want to tell.

Neither of these choices is exclusive. Indeed, often the best approach is to mix the two. As I write, Strange Company is doing preparation work on our next film, Bloodspell, which will combine live actors acting in close-up scenes with gigantic AI-controlled 50+-character battle scenes. It’s much easier to direct live actors, but I can’t fit 50 people in my office, so we use a combination approach.

Finally, you need to choose how you want to distribute your film, either in a conventional video format such as AVI or MPEG, or in a special file to play back in the game. The advantages and disadvantages here are pretty simple. More people can easily watch MPEG, AVI, or Quicktime files, but you’ll get higher resolution for smaller file size if you distribute in-game. A lot of the time, your game choice will dictate your distribution method. If your game can’t play back demos, you need to capture the action of your film using a video capture card. At that point, you can’t go back to in-game.

Enough theory. Now it’s time to choose an engine.

Does the content available in the game fit your vision? If you want to make SF action, Unreal Tournament 2004 is a pretty good bet. If you prefer a gripping noir saga, though, Max Payne’s more your style.

Evaluate all the content in the game, in context, not just in an editor (this means playing the game—bummer!), and remember to check out the available mod content. Neverwinter Nights, for example, has good default content, but the vast, vast array of available mod content really lifts it as a machinima engine. However, be aware that mod content often won’t work as well as the original game content.

That a game doesn’t have a large community around it tells you two things. First, you’ll be on your own when figuring out the intense weirdnesses of how the engine works (and all games are intensely weird and frustrating in their own ways). Second, if no one’s modding the game, there may be a reason for that; perhaps the game’s a bear to mod.

The game’s community is also your primary audience for any film. A tiny community may well predict a tiny audience.

A huge feature list isn’t the most important part of a machinima project, but some features are very useful.

Check that the game can somehow do lip-syncing. That mechanism could be skeletal animation (Half-Life), morph targets (Anachronox or Half-Life 2), or even simple skin swapping (many Quake 2 machinima). If there’s no lip-syncing in the game itself, perhaps you can either animate the heads of characters somehow or quickly swap textures on the head. Neverwinter Nights, for example, has no built-in lip-syncing, leading Strange Company to some very interesting reverse-engineering to make a movie in it.

You don’t always need lip-syncing—Red vs Blue doesn’t have it—but it looks odd to see a character’s mouth remain rigid while he’s talking.

Skeletal animation is good. Blended animation and procedural animation are better. The former means that you can combine several different animations at once; the latter means that you can directly control individual bones on the model at runtime. Blended animation is extremely useful for characters who look around as you move the mouse.

Anything that makes your movie look pretty is great. Red vs Blue succeeded in part because the Halo engine is so pretty, and in part because Halo’s procedural animation lets the actors move the characters’ heads.

A scripting language is very helpful, as is built-in machinima functionality. However, don’t expect much here. For starters, if you’re filming live, you won’t use these features much. Second, most machinima scripting systems aren’t great. There are exceptions to this rule, with cheers to both Ritual Entertainment and Epic Megagames, but even these systems need tweaking, modification, and even abandoning, depending on how you make your film.

Look for machinima scripting languages that use real film metaphors. If a scripting language talks about creating dolly shots, crane shots, and track-and-pan shots, there’s a good chance it’s pretty solid. A good scripting language, whether it’s machinima-capable or not, is a great asset to the machinima creator.

The granddaddy of them all is still alive and kicking, thanks to much open source coding. The engine lacks skeletal animation but features some pretty good graphical improvements despite its age, with quick skin swaps and tons of content and support. In particular, the machinima support for this engine is excellent; check out the Machinima Production Kit (http://www.strangecompany.org/page.php?id=50).

This engine ages like a fine wine; it’s still probably the most popular game engine ever. Automatic lip-syncing, some neat skeletal stuff, a fair degree of machinima support, and a huge community make HL a good option, despite its aged graphics. Machinima.com has some introductory tutorials (http://www.machinima.com/displayarticle2.php?article=92).

This engine is a strong contender. It has good to great support, including a lot of support from the creators, a robust machinima system, and a pretty nice-looking engine, plus lots of content. UT 2004 seems to be the choice of a new generation. See the Machinima.com tutorial (http://www.machinima.com/displayarticle2.php?article=316).

This is a slightly strange choice. It has a huge community and masses of content, but the slightly old, although still pretty, engine has some strange limitations. It has skeletal animation but no controllable procedural animation and no blending. Lip-syncing is a complete bear. I recommend it for advanced users; it’s also your best option for a fantasy film.

There’s no modding, but the engine is very, very pretty, with a huge community. The “no modding” thing, though, really does limit what you can do with it. Red vs Blue worked within these limitations, but that’s a hard trick to repeat.

It’s not quite out as I write, but everyone expects this to be the Machinima Engine of the Gods. It promises advanced physics, beautiful graphics, powerful scripting, and advanced facial animation, plus lots and lots of mod support and a huge community that will follow it over. HL2 should be the next big thing indeed. Some people predict that it will take machinima into the mainstream.