Stop Staring

Facial Modeling and Animation Done Right

Jason Osipa

• Publisher: Sybex
• Year: 2010
• ISBN:
  9780470939611
• Edition: 2
• Topic: Computer Science
• Subtopic: Digital Media
• Index: Computer Science

The de facto official source on facial animation—now updated!

If you want to do character facial modeling and animation at the high levels achieved in today's films and games, Stop Staring: Facial Modeling and Animation Done Right, Third Edition, is for you. While thoroughly covering the basics such as squash and stretch, lip syncs, and much more, this new edition has been thoroughly updated to capture the very newest professional design techniques, as well as changes in software, including using Python to automate tasks.

• Shows you how to create facial animation for movies, games, and more
• Provides in-depth techniques and tips for everyone from students and beginners to high-level professional animators and directors currently in the field
• Features the author's valuable insights from his own extensive experience in the field
• Covers the basics such as squash and stretch, color and shading, and lip syncs, as well as how to automate processes using Python

Breathe life into your creations with this important book, considered by many studio 3D artists to be the quintessential reference on facial animation.

Part I: Getting to Know the Face

Before we start animating, building, or rigging anything, let’s be sure we’re speaking the same language. In Chapter 1, I talk about talking, pointing out the things that are important in speech visually and isolating the things that are not. Narrowing our focus to lip sync gives a good base from which to build the more complicated aspects of the work later. In Chapter 2, I define and outline, in the same focused way, the top half of the face. In Chapter 3, we zoom back to the entire face—the tilt of the head, wrinkles being a good thing, and even parts of the face you didn’t know were important.

Each chapter in this part is expanded into a detailed explanation in a later part of the book: Chapter 1 in Part II, Chapter 2 in Part III, and Chapter 3 in Part IV.

Chapter 1: Learning the Basics of Lip Sync

Chapter 2: What the Eyes and Brows Tell Us

Chapter 3: Facial Landmarking

Chapter 1

Learning the Basics of Lip Sync

In modeling for facial animation, mix and match is the name of the game. Instead of building individual specialized shapes for every phoneme and expression, like for an F or a T, we’ll build shapes that are broader in their application, like wide or narrow, and use combinations of them to create all those other specialized shapes. On the animation front, it’s all about efficiency. You want to spend your time being creative and animating, not fighting with the complexities that often emerge from having a face with great range. It doesn’t sound like there’s much to these concepts for modeling and animating, and, yeah, they really are small and simple—but they’re huge in their details, so let’s get into them.

Before we can jump into re-creating the things we see and understand on faces, we need to first identify those things we see and understand. Starting on the ground floor, this chapter breaks down the essentials of lip sync. Next, we’ll go into how basic speech can be broken into two basic cycles of movement, which is what makes the sync portion of this book so simple. Finally, at the end of this chapter, we’ll take those two things—what’s essential and the two cycles—and build them into a technique for animating.

• The bare-bones essentials of lip sync
• The two speech cycles
• Starting with what’s most important: visemes
• Building the simplest sync

The Essentials of Lip Sync

People overcomplicate things. It’s easy to assume that anything that looks good must also be complex. In the world of 3D animation, where programs are packed with mile after mile of options, tools, and dialog boxes, overcomplication can be an especially easy trap to fall into. Not using every feature available to you is a good start in refining any technique in 3D, and not always using the recommended tools is when you’re really advancing and thinking outside the box. Many programs have controls and systems geared for facial animation, but you can usually find better tools for the job in their arsenals.

If you’re fairly new to 3D, and have dabbled with lip sync, it has probably been frustrating, complicated, difficult, and unrewarding. In the end, most people are just glad to be done with it and regret deciding to involve sync in their project. We’re starting to see some amazing results come from facial motion capture techniques, but at least for now, that’s probably beyond the cost range for readers of this book. Automated techniques are always improving too, but so far, they aren’t keeping up with what a good animator or capture technique can deliver.

Don’t despair. I will get you set up for the sync part of things quickly and painlessly so you can spend your time on performance (the fun stuff!). If your bag is automation, there’s still a lot of information in here you can use to bump the quality of that up too.

When teased apart properly, the lip sync portion of facial animation is the easiest to understand because it’s the simplest. You see, people’s mouths don’t do that much during speech. Things like smiles and frowns and all sorts of neat gooey faces are cool, and we’ll get to them later, but for now we’re just talking sync. Plain old speech. Deadpan and emotionless and, well, boring, is where our base will be. Now, you’re probably thinking, “Hey! My face can do all sorts of stuff! I don’t want to create boring animation!” Well, you’re right on both counts: Your face can do all sorts of things, and who really wants to do boring animation? Nobody! For the basics, however, this is a case of learning to walk before you can run. For now, we’re not going to complicate it. If we jumped right into a world with hundreds or even thousands of verbal and emotional poses (which is how they do it in the movies), we’d never get anywhere. So, to make sure you’re ready for the advanced hands-on work later, we’re focusing on the most basic concept now: bare-bones lip sync. When dealing with the essentials of lip sync and studying people, there are just two basic motions. The mouth goes Open/Closed, and it goes Wide/Narrow, as illustrated in Figure 1-1.

Figure 1-1: A human mouth in the four basic poses

At its core, that’s really all that speech entails. When lip-syncing a character with a plain circle for a mouth (which we’ll do in just a minute), the shapes in Figure 1-2 are all that’s needed to create the illusion of speech.

Figure 1-2: A circular spline mouth in the same four basic poses

Your reaction to this very short list of two motions might be, “What about poses like F where I bite my lip, or L where I roll up my tongue?” Ignoring that kind of specificity is precisely the point right now. We’re ignoring those highly specialized shapes and stripping the building blocks down to what is absolutely necessary to be understood visually. If these two ranges—from Open to Closed and Wide to Narrow—are all you have to draw on, you become creative with how to utilize them. Things like F get pared back to “sort-of closed.” When you animate this way and stop the animation on the frame where the “sort of closed” is standing in for an F, it is easy to say, “That’s not an F!” But in motion, you hardly notice the lack of the specific shape—and motion is what I’m really talking about here. You should be less concerned with the individual frames and more concerned with the motion and the impression that it creates. For most animators, there is a strong instinct to add more and more complexity too early in the lip-sync process, but too much detail in the sync can actually detract from the acting.

Animating lip sync is all illusion. What would really be happening isn’t nearly as relevant as the impression of what is happening. How about M? You may be thinking, “I need to roll my lips in together to say M, and I can’t do that with a wide-narrow-mouth-thingamajig.” Sure you can, or at least you can give the impression in motion that the lips are rolled in—just close the mouth all the way—and that’s usually going to be good enough. When you get the lip sync good enough to create an impression of speech and then focus your energies on the acting, others will also focus on the acting, which is precisely what you want them to do.

Analyzing the Right Things

Let me take you on a small real-world tutorial of what is and what is not important in speech.

Animators have a tendency to slow things down to a super-slow-mo or frame-by-frame level and analyze in excruciating detail what happens so as to re-create it. This is not necessarily a bad thing, but here’s an example of how that can break down as a method: Look in the mirror, and then slowly and deliberately overenunciate the word pebble: PEH-BULL. You’re trying to see exactly what happens with your face. Watch all the details of what your lips are doing: the little puff in your cheeks after the B; the way the pursing of your lips for P is different than for B; how your tongue starts its way to the roof of your mouth early in the B sound and stays there until just a split second after the end of the word. You’d think that all these details give you a better idea ...