Monday, January 14, 2008

The Creator character

The puppet is being constructed as a marionette. This approach allows for a rich and varied kinematics1 and dynamic behavior2. Mechanically it resembles a cable-operated rigid body (such as a complex crane). The puppet consists of a series of joined rigid elements, which correspond to the simplified mechanism of a human body.

The non-verbal language of the puppet is constructed as a neural net of three simple primitives that control gestures (limbs, hands), speech (the puppet “vocal chords”), and expressions (body language).3

The essential role of these primitives is to build a vocabulary of behaviors4. The combination of these primitives can give rise to more gestures with different meanings and evolve complex behaviors by sequencing, superposition and inhibition5.

The mechanical design consists of determining the number of limbs, type of joints, degrees of freedom (DOF)6 of the joints and other movable pieces. It is not defined yet how many strings will be used to control the marionette, but probably between 8 to 12 will be enough, depending on various factors including number of segments of limbs and cost and time of development for both software and hardware.

1- In physics, kinematics is the branch of classical mechanics concerned with describing the motions of objects without considering the factors that cause or affect the motion. By contrast, the science of dynamics is concerned with the forces and interactions that produce or affect the motion. The term "kinematics" derives from the Greek word ______, meaning "to move".

2- The basis of kinematics is the choice of coordinates that describe the position(s) and/or orientation(s) of object(s). The time derivatives of these coordinates correspond to velocities and accelerations. An important component of kinematics is differentiating position to obtain the velocity or acceleration, and vice versa, integrating velocity or acceleration to obtain the position. Another key component of kinematics is converting between different sets of coordinates that describe the same motion. Both of these components are fundamentally mathematical methods, and are not based on physical principles.

The simplest form of kinematics is the description of point particle motion (translational kinematics). The description of rotation (rotational kinematics) is more complicated. The state of a generic rigid body may be described by combining both translational and rotational kinematics (rigid-body kinematics). The most complicated case is the kinematics of a system of rigid bodies.

In all of these cases, the most useful choice of coordinates may be determined by constraints on the motion, or by the geometrical nature of the force causing or affecting the motion. For example, it may be convenient to describe the motion of a bead constrained on a circular hoop by its angle on the hoop. Similarly, it may be practical for calculations to describe the motion of a particle acted upon by a central force using polar coordinates. (Source: Wikipedia)

3- Emotions as expressed by “matchstick” body language.

The question here is: how much visual information is needed to express character? Even though I have done motion capture sessions a number of times, it does not cease to amaze me the amount of “personal” information that is actually registered by recording the markers placed on our basic joints. Without the benefit of face expression, body shape (other than general proportions) and other elements that would seem to convey a great deal about a subject, the identity and some key “personality traits” can be immediately deduced or understood by the visual combination of a set of lines in motion.

Here is an example. Years ago I had to capture the movements of the entire team of the “A’s” of Oakland, then World Baseball Champions. I used a recent invention by a small startup called BioVision which used 4 cameras to record the position of reflective markers set at the joints of the player’s limbs as well as other locations like the chest, shoulders etc.

Reflective markers for motion capture (Source: Brunel University, London)

The position of these points of light were correlated with an initial calibrating position and sent as data to the program, which triangulated and calculated the 3d position of such markers. These positions were then transferred to a stick figure consisting of the same number of limbs and joints. The movement could then be played back. At the sight of the stick figures moving, my coworkers, who were all baseball fans, were able to tell immediately the identity of the character long before it made a move that gave away his role. Sometimes just the stance, the gait or a slight movement of the “body” would carry the personality, which was familiar to the observers.

Motion capture data applied to stick figures (Source: University of California, Irvine)

This experience made me think about what it is that constitutes that personality. How do we express emotions and how do we “read” other people? It seemed that by stripping away all the particular characteristics of the physical body except the basic proportions and movement (time-displacement or space-time of the body) we could convey the story in an even more direct way. The empathy and identification was, in fact, enhanced. By establishing a direct connection with some part of our psyche or perhaps our animal brain, it gave new meaning to the phrase “judge of character”

4- McNeill, David, Language and Gesture, Cambridge, University Press, 2000

5- According to David McNeill there are four kinds of gestures that are usually distinguished:
Symbolic gestures represent some widely recognized conventionalized meanings with body configurations and movements. For example, a thumbs-up gesture indicates agreement.
Deictic gestures are those that point objects in the environment with arms and figures. They may also be used to indicate unseen, abstract or imaginary things.
Iconic gestures are some pictorial gestures intuitively representing physical entities in the world.
Metaphoric gestures represent abstract objects or concepts.

6- In mechanics, degrees of freedom (DOF) are the set of independent displacements that specify completely the displaced or deformed position of the body or system. This is a fundamental concept relating to systems of moving bodies in mechanical engineering, aeronautical engineering, robotics, structural engineering, etc.

A particle that moves in three dimensional space has three translational displacement components as DOFs, while a rigid body would have at most six DOFs including three rotations. Translation is the ability to move without rotating, while rotation is an angular motions about some axis.

No comments: