General Overview
General Overview
Why physics ?
• Because most things in our everyday
environment can be described by physics
–and a common ambition in a simulation
is to describe an environment
• Physical simulations capture complexity
• Complex behavior emerges from simulation
General Overview
Luxo Jr. by John Lasseter, Pixar 1987
• Keyframed realistic animation with no
physical simulation.
• Still, physics does the trick…
• High level motion control: Jump from A to B
• Luxo Jr. made people start thinking about physics . . .
General Overview
Physics is . . .
• Systematic
• Scalable
• Consequent
• Controllable
• Extensible
• General
Therefore library software and expertise can evolve !
General Overview
Applying Physics can create . . .
• Dynamics (Newton’s Laws)
• Rigid Body Dynamics, Interaction, Collisions
• Mechanics of Materials
• Fluids, water (water, yacht)
• Gases, smoke, clouds (smoke, train, interactive, clouds)
• Plasmas, fire, lightning, sparks (candle)
• Particles (special effects, explosions, fountain, hair, fur)
• Fields
• Body interaction
• Collective phenomena, complex systems, emergence, …
• Audio – wave tracing, acoustics, damping, …
• Light – optics, ray tracing, …
• Effectors and sensors – display and interaction
In the Beginning . . .
In the Beginning . . .
• Computers were hulking Goliaths locked in air-conditioned
rooms.
• But a young electrical engineer and former naval radar
technician named Douglas Engelbart viewed them differently.
• Engelbart envisioned them as tools for digital display.
• He knew from his days with radar that
any digital information could be viewed
on a screen. Why not, he then reasoned,
connect the computer to a screen and
use both to solve problems?
In the Beginning . . .
• Engelbart's ideas were dismissed, but by the early 1960s other
people were thinking the same way.
• Moreover, the time was right for his vision of computing.
Communications technology was intersecting with computing
and graphics technology.
• This synergy yielded more user-friendly
computers, which laid the groundwork for
personal computers, computer graphics,
and later simulations.
In the Beginning . . .
Pivotal Points in History . . .
Fear of nuclear attack prompted the U.S. military to commission
a new radar system that would process large amounts of
information and immediately display it in a form that humans
could readily understand. The resulting radar defense system was
the first "real time," or instantaneous, simulation of data.
• Aircraft designers began experimenting with ways for computers
to graphically display, or model, air flow data.
• Computer experts began restructuring computers so they would
display these models as well as compute them. The designers'
work paved the way for scientific visualization, an advanced form
of computer modeling that expresses multiple sets of data as
images and simulations.
In the Beginning . . .
More Pivotal Points in History . . .
• One of the most influential antecedents of today’s simulations
was the flight simulator. Following World War II and through the
1990s, the military and industrial complex pumped millions of
dollars into technology to simulate flying airplanes (and later
driving tanks and steering ships).
• By the 1970s, computer-generated graphics
had replaced videos and models. These
flight simulations were operating in
real time, though the graphics were
primitive. By the early 1980s, better
software, hardware, and motion-control
platforms enabled pilots to navigate
through highly detailed virtual worlds.
In the Beginning . . .
Even More Pivotal Points History . . .
• Of course, the "military-industrial complex" was not the
only entity interested in computer graphics.
•A natural consumer of computer graphics was the
entertainment industry, which, like the military and industry,
was the source of many valuable spin-offs in virtual reality.
• By the 1970s, some of Hollywood's most dazzling special
effects were computer-generated, such as the battle scenes
in the big-budget, blockbuster science fiction movie Star
Wars, which was released in 1976. Later came such movies
as Terminator and Jurassic Park. In the early 1980s, the
video game business boomed.
Engineering Techniques and Feats in
Physics Simulation
Current Climate
• Physics-based simulation is the art of reducing algorithmic
complexity to achieve usable results with as little computation as
possible.
• Current visualizations usually incorporate both scientific
simulations based on research data and parametric rule-sets that
augment the data and save time.
• State of the art simulation technology in this field is most often
proprietary and created by specialized teams (scientists, engineers,
artists, specialists) when the need for them arises.
• A large amount of cooperation between research institutions,
productions houses, and software developers is needed and
planning for computational demands is highly coordinated.
Current Climate
Some General Approaches Used:
• System Modeling (pre-computed) – Simulation based on
equations (x,y,z,t) that are computed for individual data points (the
more the better).
• Parametric Rule-Sets (pre-computed) – Integrates pre-
computed systems with other simulations (from artists or
augmented and reduced systems) based on defined parameters.
• Selective Data Plotting (real-time) – Compute only the data
points you need to get an accurate representation of your system.
Example: Ray-casting
• Monte Carlo Simulation (depends on application) – Compute
complex systems using algorithms with reduced complexity by
guessing (using known boundaries).
Current Climate
Areas Where Simulations Are Used in the Industry:
• Newtonian Physics - Rigid Body Kinetics, Object
Collisions, etc.
• Realistic Human Movement – Muscle Behavior, Skeleton
Behavior, etc.
• Volumetric Rendering – Natural Phenomenon (fire,
smoke, clouds, fluids, solid bodies)
• Large-Scale Simulations – Flocking, Group Behavior, and
Character Interaction
Newtonian Physics
• Newest
implementation
allows user control
over bodies.
• Basic laws of Newtonian physics are readily defined
(ex: a(t) = v’(t) = s’’(t)).
• Uses algorithms with interpolate the appropriate physical
simulation based on user-defined actions (similar to key-framing).
• Current hurdles involve number of bodies, which can interact in
real-time, and how many parameters are checked for each instance
of time.
• Research at Carnegie Mellon University and implementation using
controller packages like Maya, Houdini, and Softimage.
Realistic Human Movement
• A new technique is being developed at the University of Toronto
that creates a composable controller system for human movement.
• It links motor abilities of characters to physics-based controllers.
• This is different from traditional character animation because
character responds to interactions based on laws of physics.
• This is done through the groups of linked controllers that contain
physical parameters.
Volumetric Rendering
• This technique has been too computationally intensive (usually
N^3 complexity) in the past.
• In the last couple of years new technologies and techniques have
arisen to reduce the complexity.
• Ray-casting – This method computes only the portions of the
volume that are seen given the current projection.
• Volumetric models of systems are
important because they are
easily integrated with both
pre-computed system model data
and parametric rule-sets to create
realistic behavior.
Volumetric Technologies
• Arete Digital Nature tools used in Cinema. Creates volumetric
fluids, clouds, etc.
• Real-time volumetric renderings can still not be computed
accurately on standard computer hardware without an additional
accelerator.
• Mitsubishi Electronics RTViz group produced to VolumePro 500
and VolumePro 1000 to do ray-casting in real-time.
• Uses a large frame buffer and specialized hardware
to compute 2563 individual volumes at 30 fps.
• Nvidia developed a technology (VTC) that
compresses 3D textures (representing slices of a
volumetric object).
Large Scale Simulation
Particles and Flocking:
• Uses a large number of discrete objects (with physical parameters
and or simulation data attached) to model systems in which many
bodies are interacting non-uniformly.
• Examples are crowds, swarms, and flocks, which could represent
characters or environmental phenomenon.
• Large-scale simulations are usually coupled with other
technologies such a volumetric rendering and Newtonian physics.
• Weta’s Massive is a very advanced, state of the art
implementation of large-scale simulation.
Current Feats
LOTR Two Towers:
• Weta used its proprietary software
Massive for large-scale simulation.
• It integrated Massive with its own
physics-based muscle-system using
Maya to create effects that would take
460 years to be rendered on a home PC.
• Rendering and simulation took 10 months and was done using
about 1000 IBM and SGI workstations.
• Their own rendering and shading program Grunt was used with
Massive to aid in simulating cloth and hair on individual characters
in the crowd.
Recent Feats
A Perfect Storm:
• First film where ILM feels they created realistic fluid
dynamics.
• Until recently, fluid flow simulations could only be run
with 80-100 data points. In “A Perfect Storm”
simulations of several different oceans were run in 3D
using a much larger data set and various parameters
until they looked right.
• Several simulation technologies went into recreating
the ocean.
Simulating Physics in the Art World
Simulation for the Artist
Ability to manipulate/distort
physics of their digital environment
Game Designer Perspective:
• Games allow Player to express himself
• Realism is not necessarily the goal for games
• Simulated World that has “Consistently”
Simulation for the Artist
Grenade example: grenade bounces realistically
Simulation for the Artist
Pro’s of digital simulation:
• Time Saved & Reduced costs
• Emergence – new strategies and
game play by user
• Simulated World that has “Consistency”
• New and more genres of games created
Simulation for the Artist
Con’s of digital simulation:
• Unexpected methods of game play by player
• Being too complex or real for the game
• More user feedback is required
• Hardware limitations reduces use or ability
• Deeper simulation does not mean more fun.
Maybe just more interesting
Simulation for the Artist
Thief by Looking Glass Studios:
Deeper awareness
model with
complex sound
propagation and
lighting used for
stimuli of
characters
Simulation for the Artist
Thief by Looking Glass Studios:
Sound simulated
to bounce off
surfaces and
materials with
varied intensity
and reverb.
Simulation for the Artist
Thief by Looking Glass Studios:
Designers added a “light gem” feedback device
To help the player utilize environments effectively
Simulation for the Artist
Simulation to the non-Game Artist:
3D programs like Maya 3D
incorporate more simulation
for manipulating and controlling
physics in 3D.
Simulation for the Artist
Simulation to the non-Game Artist:
KPT makes Adobe Photoshop plug-ins
utilize 3D physics simulation of light &
movement upon 2D surfaces.
Ex.
Goo
Gel
Materizlier
Turbulence
Goo plug-in original
What Simulation means to the Public
What the General Public will see:
• Simulation-based game design produce
more variable player driven game play
• More dynamic effects in movies
produced with digital tools
Movie Industry Example:
LOTR Used
motion-capturing
to create cycles
to be used with
AI and physics to
simulate mass
warfare.
Future of Virtual
Simulation and Visualization
Future of Virtual Simulation and Visualization
• Current problems that keep our technology
from advancing
• Overcoming technological barriers and what
to expect from future simulation
• Real-time simulation and visualization
• Advancements in representing human
figures in motion
• Realistic walking movement
• Physics-based human simulation for virtual prototyping
• Human populations in simulations
• Technology that can be used for motion planning in
robots
Future of Virtual Simulation and Visualization
Problems that keep our
Simulation technology from advancing
Today’s simulations require intense
computational resources
• Most simulations generate gigabytes to terabytes of data,
whether it is a scientific application or an artistic application such
as movie effects
• This Data requires not only huge amounts of storage space but
also the computer processing power to handle it.
Future of Virtual Simulation and Visualization
Example of a movie simulation: Water
When simulating water, dynamics and large
scale 3D grids are used. Each point on the
grid can generate thousands to millions of
particles that are tracked over time.
Future of Virtual Simulation and Visualization
Example of a movie simulation: Water
The amount of data generated is so immense
that in many cases artists must manually add
in details because the movie production can
not afford so much simulation time.
Future of Virtual Simulation and Visualization
Overcoming technological barriers and what
to expect from future simulation
As computers get faster and storage space
becomes cheaper, more complex and realistic
visualizations are possible
Real time visualization will revolutionize the
world of simulation.
Future of Virtual Simulation and Visualization
Real time means that the graphical outcome
of a simulation will be available as the
computer works through the simulation.
What does this mean for artists?
Artists will be able to create very complex
visuals without waiting to see the final
product. Changes will be possible without the
fear of long rendering times.
Future of Virtual Simulation and Visualization
What does this mean for scientists?
Scientists will be able to simultaneously get
data from the simulation while analyzing the
computer visualization.
Future of Virtual Simulation and Visualization
How far have we come with real time?
Pixar’s first film, “Luxo, Jr.” was made almost 17 years ago. The
short film was rendered on a Cray supercomputer that took 75
hours per second of animation.
Future animated
films could be done entirely
on desktop machines as
hardware advances.
Future of Virtual Simulation and Visualization
Advancements in representing human
figures in motion
Research is being done to develop systems to more
realistically display human movement such as walking.
Today’s systems pass for situations not intended to be
“realistic” but when actual human-like figures are represented,
we can notice that there is something not right.
Physics are what control human movement. The combination
of physics and other techniques such as motion capturing can
yield very realistic human motion. There is still a lot to hope
for.
Future of Virtual Simulation and Visualization
Final Fantasy uses CG to represent Humans
Square’s Final
Fantasy: The Spirits
Within raised the bar
for technical
achievement as it
made digital actors
look realistic.
Characters moved almost too gracefully as real humans tend to
be somewhat more jerky and unpredictable. We can expect to
see much more realistic movement in future CG movies.
Future of Virtual Simulation and Visualization
Future Applications of Realistic Human Motion
Physics-based Human Simulation
for Virtual Prototyping
Boston Dynamics has been developing
a landmark 3D software product
for real time simulations called
DI-Guy. This software adds artificial
human life to simulations such as
virtual battlefields.
Future of Virtual Simulation and Visualization
Future Applications of Realistic Human Motion
Technology for Virtual Humans could provide potential for
humanoid robots.
Robots use complex physics to perform
simple motion. In the next 10 years,
Humanoid robots could be a common
sight. The development of these new
robots could be furthered by the same
research done to represent human
movement graphically.