Monday, November 26, 2007

Madame Tutli Putli & Performance Capture

Continuing on the subject of the threshold between having character and being a character (as well as the emotional connection that determines what we perceive as a character), Madame Tutli Putli is a really neat example of a mixed media character design. This is a stop-motion film, but the eyes of the main character were composited in from footage of a live actor who was recorded after the animation. The result is a very unique, often creepy, look.

This process is a unique form of performance capture in which the performance occurs only after the animation, and is probably why this film has avoided the controversy that typically surrounds optical motion capture such as Beowulf.

But in a sense, Madame Tutli Putli is much closer to reality than optical motion capture, simply because there are no markers, but the actual actor footage is used. Moreover, the eyes are arguably the most important component of facial animation. Certainly, the footage is used very creatively, but there are a lot more adjustments and tweaks to the data in the optical mocap pipeline.

One key difference lies in the role of the animator. In Madame Tutli Putli, the animator's performance drives the actor's contribution, whereas in typical mocap productions, it is the actor's performance that is most important. A subtle difference, but one that appears to be central to the mocap debate.

Friday, November 23, 2007

More thoughts on what makes a character

Can water or smoke be a character? In the era of big-budget effects films, it's common to hear people refer to "hero" waves or drops or puffs. So are these "hero" elements characters as well?

Animators are taught from the very beginning to exaggerate what they see in real life, an idea that is manifest in various principles such as squash-and-stretch, anticipation, breaking joints, etc. But even though an animation as simple a bouncing ball can have character, this does not mean that a bouncing ball IS a character. To give an animated object character is to imbue it with personality, but a character is more than just personality - it is an entity that we can recognize as a living, thinking being i.e. one of our own.

That is not to say that a bouncing ball can never be a character (there are obviously shades of grey here), but I feel that there is a threshold that separates the bouncing ball with character from the bouncing ball that is a character.

Phew, that was a real tongue-twister of a post.

The first few seconds of this video also depicts the difference between having character and being a character. The water has personality, but then morphs into a water character.

Wednesday, November 14, 2007

Automated Anticipation?

I came across a paper entitled "Anticipation Effect Generation for Character Animation". Basically, these researchers were looking at automated ways to add anticipation to existing animation (presumably 3d animation). It's an interesting notion, and seems to tackle the subtleties of animation from the opposite perspective as motion capture. This technique aims to make it easier to improve keyframe animation. Most animators would object to this approach on the grounds that anticipation is subjective and not easily derived automatically. The researchers found the same thing - they could not automate the duration of the anticipation and needed human intervention.

Here's the abstract:
"According to the principles of traditional 2D animation techniques, anticipation makes an animation convincing and expressive. In this paper, we present a method to generate anticipation effects for an existing animation. The proposed method is based on the visual characteristics of anticipation, that is, “Before we go one way, first we go the other way [1].” We first analyze the rotation of each joint and the movement of the center of mass during a given action, where the anticipation effects are added. Reversing the directions of rotation and translation, we can obtain an initially guessed anticipatory pose. By means of a nonlinear optimization technique, we can obtain a consequent anticipatory pose to place the center of mass at a proper location. Finally, we can generate the anticipation effects by compositing the anticipatory pose with a given action, while considering the continuity at junction and preserving the high frequency components of the given action. Experimental results show that the proposed method can produce the anticipatory pose successfully and quickly, and generate convincing and expressive anticipation effects."

The entire paper can be found at: and the pdf is at

(You may need to be on USC campus to be able to access the links)

Tuesday, November 13, 2007

Beowulf's Cousin

This is an example of motion capture that I did this semester at USC using a live performer.

The clip illustrates two things: first, mocap can produce very subtle and natural movement when it comes to body motions. Second, as demonstrated by the problems with the character's arm when he falls down, motion capture is prone to problems due to occlusion (i.e. the markers are not visible by enough cameras), and data needs to be manually reconstructed by an animator.

Character Research

Here is some of my own work that examines what constitutes a character - in this case, it's a bridge.

Tuesday, November 6, 2007