标题: Skeleton-based diverse creature design tool for mass production [打印本页] 作者: 彬彬 时间: 2011-12-29 08:54 标题: Skeleton-based diverse creature design tool for mass production 1 Motivation
The purpose of our project is to develop digital tools that will help
to simplify the task of creature design in large-scale media projects.
When working on an animation or game that requires hundreds of
unique creature designs, many artists find it challenging to maintain
consistency and quality in every new design. As a result, even the
most creative artists must often go through a long and tedious period
of revisions before they arrive at final creature designs that are
satisfactory to the director. To reduce the amount of time and resources
needed for this critical revisions stage, we have designed a
prototyping tool which can use existing designs to rapidly generate
unique iterations.
2 Backgrounds
There have been attempts to create prototyping tools in the past, but
with limited success. Asset management software programs such as
Shotgun have prototyping capabilities, but are not integrated into
standard workbenches such as Maya and thus are not part of most
artists’ workflows. Genetic Algorithm (GA) based creature exploration
tools have also been used for generating new designs automatically,
but this software allows for an infinite number of possible
combinations, requiring substantial human input to focus its
designs to match the originals. In response, we have designed our
software to interface more seamlessly with existing animation tools,
as well as simplified its functionality to make it more intuitive for
artists.
3 Current Workflow
The One Art Director then reviews the new designs to determine if
they match his or her original concept, and will ask for revisions
when necessary. This is a vital step, because it is often the Art Director’s
last opportunity to make sure that all the creature designs
are aesthetically consistent with each other and with the environment
of the game or film. After the designs are approved, they are
modeled, textured, and rigged for animation and final rendering. At
this stage, it is difficult for the Art Director to modify unsatisfactory
designs. As such, the Art Director will sometimes restrict character
designs from passing through the 3d production pipeline, until all
the designs for the characters have been finalized first as concept
drawings.
e-mail: shirai@mail.com
4 Skeleton-based diverse creature design
We propose a software solution which will speed the iterative process
by designing creature skeletal s***ctures before sketches and
mesh modeling. The advantage to this new order of stages is that
by modifying a simple skeleton, artists are able to create and judge
new designs semi-automatically much more quickly than if they
must fully re-model the character for each iterations. This faster
work-flow can also be used to test less conventional creature designs,
such as radical monsters or mechanical designs that have no
predecessors in existing media. This tool is s***ctured by 13 buttons
of template s***cture, 1 scale slider and 4 relation controllers which
is written in 319 lines of MEL (Maya Embedded Language) script.
Artists can load this MEL tool in earlier stage of their workflow to
explore their own creature. It based on template s***cture like Human
body, head, arm-L, arm-R, leg-L and leg-R or Dragon body,
head, wing-L, wing-R, leg-L, leg-R, tail. Artists can add body parts
by clicking a button and change their scale and relations by other
controllers. They can explore the character design by adding, removing
and moving body parts in 3D view. This tool can be ignored
and it has a compatibility in current workflow.
5 Result and Conclusion
User testing provided essential feedback during the development of
our software, indicating both its successes and possible improvements
for later iterations. In our most recent tests, an artist who
can create tarot 22 creatures in 20 days in sketch based workflow
was able to create 10 base creatures in 3 days with skeletal s***ctures
and sketches. Evaluating artists also found the value of our
tool, it can expand creature diversity with biological growth and
it can also useful to avoid conflicting with past similar creations.
In future versions of this software, we hope to make the graphical
user interface (GUI) more s***ctured and intuitive, and to include
more creature types, such as mollusks, insects, and plants as additional
design sets. We also hope to better integrate our tools with
physics simulations and artificial intelligence engines for increased
compatibility with other programs.
Our goal is to streamline the workflow of professional production
pipelines through the design and implementation of innovative digital
tools. Interacting with professional designers has informed us
of the challenges they face in their everyday work, and helped us refine
and focus our software to become a powerful imagination aide