Computer Animation Techniques: Keyframing, Interpolation, and Physical Simulation, Lab Reports of Computer Science

Download Computer Animation Techniques: Keyframing, Interpolation, and Physical Simulation and more Lab Reports Computer Science in PDF only on Docsity!

1

Computer Animation Techniques

Goals

ƒ Introduce 3 forms of animation

„ Keyframing

„ Simulation

„ Motion Capture

3

Keyframing- Traditional 2D

ƒ Highly skilled animator draws the

important, or key frames at the desired times

ƒ Less skilled (lower paid) animator draws

the in-between frames - 1 picture for every 1/30 th^ of a second

Keyframing in 3D

ƒ Animator specifies the important key

frames

ƒ Computer generates the in-betweens

automatically by interpolating

7

Big Picture

0 0.2 0.4 0.6 0.8^ t 1

1

0

s

speed curve (^) space curve

= animator control

t = 0

t = 1

t = 2 t = 3

t = 6 t = 8

Linear Interpolation

t = 0

t = 1

t = 2

t = 3

t = 8 t = 6

1 0

0

0 u u x x

u u

x u x −

where t 0 is the parameter value at the start of a segment, t 1 is the parameter value at the end of a segment and t is the parameter value for which you want to find the position, x intermediate x

start key x

total x distance between start and end “a portion of” key

9

Cubic Curve Interpolation

ƒ Cubic curve that is “bent” to go through

all points

t = 0

t = 1

t = 2 t = 3

t = 8 t = 6

cubic curve interpolation

Tangent Based Control

ƒ Another way to control speed and

space at same time

ƒ Used in Maya

13

Physical Simulation

ƒ Physical Simulation

„ Use the laws of physics to determine the

“in-betweens”

„ specify start conditions

„ equations of motion do the rest

Physical Simulation

ƒ Pros

„ realistic motion

„ good for interactivity

ƒ Cons

„ computationally expensive

„ difficult to control

„ little room for artistic expression

15

Physical Simulation

James O’Brien

Wayne Wooten

Passive

Active

Passive Physical Systems

ƒ No sources of energy – muscles,

motors, etc

User

initial conditions Model

Numerical Integrator

state(x,y,z,R) Graphics Engine water spray leaves kite clothing

19

Point Mass Physics

We know exactly how a point mass particle behaves

m

f

a

f ma

v ( t + ∆ t )= v ( t )+ a ( t )∆ t

2 () 2

1 x ( t +∆ t )= x ( t )+ v ( t )∆ t + a t ∆ t

Point Mass Euler Step

Euler_Step(pos, vel, f, float dt) { acc = f/m; newVel = vel + accdt; newPos = pos + veldt + 0.5accdt*dt;

}

Not Generic – specific to a point mass!

21

Particle System Recipe

ƒ Clear forces

ƒ Compute New Forces

ƒ Take an Euler Step to get new position,

new velocity

ƒ Update State

ƒ Draw Particles

ƒ Repeat

Particle Systems

ƒ Sims – Particle Dreams

25

Motion Capture Applications

ƒ Animation

ƒ Interactive characters

ƒ Robot control

27

Motion Capture Approach

ƒ track motion of reference points ƒ convert to joint angles ƒ use angles to drive an articulated 3-D model ƒ motion paths can then be adapted and generalized

Why not use the reference points directly?

How to use the data?

ƒ Off-line

„ Processing by filtering, inverse

kinematics

„ Produce libraries of motion

trajectories

Š Choose among them Š Blend between them Š Modify on the fly

ƒ On-line (performance animation)

„ Driving character directly based on

what actor does in real time

31

What can be captured?

dynamic or slow moving?

Titanic, House of Moves

What can be captured?

large scale small scale

Motion Analysis

33

What can be captured?

"rigid" body

motion

flexible objects

Titanic, House of Moves

What can be captured?

ƒ props often cause

problems

„ ball in ping pong

„ fly fishing

„ sword

ƒ passive behaviors are

hard

„ complicated motions of

clothes

„ explosions & water?

37

Ascension, Polhemus

Position and orientation

Technologies: Magnetic

Technologies: Magnetic

heavier sensors (more flop) wires on body (wireless back to base station) both position and orientation information real time $60K ($2K/additional marker) limited accuracy (~10x less accuracy than optical) smaller workspace spikes in data -> filtering ~80 hz max sensors are the cost and so it doesn't scale sensitive to EMI/ metal, particularly in floor— hard to debug

39

Technologies: Exoskeleton

Analogous, Sarcos some restriction of movement assumptions of transformation to rigid body motion made at time of design of system another technology needed for the root node not range limited high frequency (500 Hz) truly real-time

Technologies: Monkey

Puppeteering of animated characters

Not exactly motion capture but exoskeleton without the person?