Table of Contents:

1. Introduction
2. After Effects/Premiere Pro with Plugins
3. After Effects/Premiere Pro “standalone”
4. Other Pro Tools
5. Free Tools

INTRODUCTION

I’ve been asked many times how to best achieve a film-quality look in machinima postproduction.

You’ve probably noticed that some machinima is excessively clean and crisp looking. Raw footage from a video game or virtual world such as Second Life can be cartoony, and not exactly cinematic. The real world is imperfect and sometimes gritty. We work in postproduction to simulate the look and feel of reality, emphasize certain elements and de-emphasize others, and convey mood, feeling and atmosphere.

Let’s take this “BEFORE” shot captured in Second Life for “Scent Memory” as an example. Despite the detailed textures, it still looks artificially crisp. The colors also appear flat, and nothing really pops out at you.

As you can see, we succeeded in the “AFTER” shot in achieving 3 important things:

1) A cinematic, authentic film look

2) Emphasis placed on certain elements without loss of detail. In fact, the character’s face (especially his eyes) benefit greatly and the viewer’s eyes are really drawn to them. The ominous appearance of the background elements remain, while their presence is downplayed to the lead character’s benefit

3) Mood. An atmospheric, brooding feel of mystery. Postprocessing can be used to add “feeling” to any shot, from “bright and cheery” to “dark and dreary,” from comedic to horrific.

In this series, I’ll list four of the ways I know starting with my preferred method.

NEXT PART >>

Great post by Dennis Jerz over at Seton Hill University that covers a variety of topics relevant to machinima. He summarizes the concept of uncanny valley in a pretty clear and straightforward way:

When animation is too real for our brain to process it as a cartoon, but not real enough for us to accept it on a deep, instinctive level; something seems wrong. We’re trained to sense trouble when someone won’t make eye contact with us, when their facial expression doesn’t match their words, when they moving stiffly (perhaps because they’ve been injured by a hazard we haven’t noticed yet).

If part of you is just a little creeped out by clowns, or china dolls, or you’re amused by zombies and other undead, it’s because those human-like creations fall along a line. To some extent, the more human-like something is, the better we respond to it. But at some point, when something looks very much like a normal human, but does not act like a normal human, our comfort level drops.

“Uncanny valley” is an easy trap to fall into with machinima production. While we now have the relative “luxuries” of realitime rendering and superfast development/production time, this is truly one of the challenges of this growing new field. Examples of uncanny valley abound in machinima, in fact, you’re more likely to find it then not if you watch a random machinima movie on YouTube.

How do we avoid it? I can’t say we will always be entirely successful at this, but we do try. Some techniques we employ are creative use of animations for purposes other than their intended ones, time remapping with After Effects, sparing use of closeup shots of avatars, and tons of tricky editing!

We also always “test screen” our footage with people NOT familiar with video games, Second Life or machinima. It’s crucial to listen to the feedback from “normal people” who have some perspective and distance from your independent machinima project, they are the ones who will clearly and unequivocally tell you if something in your production is “creepy,” “off,” etc. Sometimes when you’re immersed in your project, your suspension of disbelief far exceeds that of the average first-time viewer.

On the other hand, if “creepy” is what you want to achieve (for instance in a horror/thriller film), then it’s easy with machinima! I’m surprised we don’t see more of these genres, given the few “advantages” of uncanny valley.

Here’s my own handrawn uncanny valley diagram:

More today on Adobe After Effects CS5‘s Roto Brush feature! In an extensive review published today at TechWorld, Jeff Foster writes…

A simple change on one frame may affect 10 to 15 frames after it, so it’s best to create short sections of about 15 frames instead of a full footage sequence from the beginning.

Sounds to me like the workflow is similar to motion tracking in After Effects CS4 and previous versions. That is, it can be somewhat time-consuming to correct, but tolerable (and even, *almost*, fun). This is in contrast to rotoscoping as I’ve always known it–a tedious, frame-by-frame nightmare that goes on for hours and seems to never end. It *is* still rotoscoping (as Michael Coleman at Adobe bluntly pointed out to me), but the difference here is huge… mainly because it can now be done in a fraction of the time! That’s precisely what has always been so nightmarish about rotoscoping–the sheer amount of time it takes.

Foster went on to say in his review…

Also, don’t expect the same quality matte results that you would get from shooting your subject against a green screen.

Clearly! Keying is always ideal if you can swing it. Rotoscoping is to be avoided at all costs (even though it’s sometimes unavoidable). Machinima in Second Life actually affords us optimal conditions for green screen/blue screen compositing, such as this depth of field effect from “Smell Memory:”

I thought about using rotoscoping for that shot, for about 5 seconds, before we had the blue screen idea. Once we’re done with the film, I’ll post a full tutorial here comparing techniques for getting this done right. Only after I install CS5, that is