One of the best parts of Imagina is the research conferences where scientists from around the globe present their latest work, sometimes in a sneak preview of their SIGGRAPH presentation. This year the focus was mathematics and simulations.

The use of procedural technology radically changes the way images, and therefore vfx, can be made, because of its auto-generative approach. When some basic parameters are set, the procedural software can automatically and endlessly generate content, thus reducing the technical workload and allowing the user to concentrate on the creative side of things. This was obvious in two very different presentations, one from Patrick Mueller on automated creation of urban environments and one from Frank Vitz, on automated character animation.

What makes vfx believable is the highly intricate level of details needed to recreate realistic images of the real world. We generally tend to assume that this cannot be achieved just by playing with sliders and parameters, perhaps because this approach gave birth to software like Poser or Bryce, that have been geared mainly towards the general public. But Patrick's presentation -- as well as Frank's -- showed that with cleverly designed sets of parameters and algorithms, one can recreate the complex patterns of nature or man-made products. There have been many movies with large cityscape environments in the past few years. Recreating and detailing such complex works is a daunting task but Patrick's work opens up new perspectives for the creation and customization of these environments through highly automated processes, not only at the city planning and implementation level, but also for the creation of each individual building in the cityscape, with the ability to automatically change the architectural style, the size and the shape of the building with the simple move of a slider. No more endless hours of modeling and remodeling!

Starting out as a visual for MTV Asia, Codehunters (U.K.), is a stunning cartoon-rendered short film directed by Ben Hibon and post-produced by Axis Animation and the Mill; it won Best Short Film at Imagina.

On the other end of the spectrum, the character behavioral simulation tools showcased by Frank Vitz were also stunning and yet, so simple. Frank is senior art director at Electronic Arts and his team is well-known for having ported the Universal Capture technology developed for the Super Punch sequence in Matrix 3 to a real-time video game engine. But, as said Frank, "This technique was not yet effective enough to cross the Uncanny Valley, the space where a human character is highly believable yet extremely zombie-like", such as seen in Polar Express or Final Fantasy: The Spirits Within. So, in order to give life to ultra real characters, Frank's team, EA's Worldwide Visualization Group, worked on isolating a series of micro-level behaviors like eye or head movements, body and facial expressions, etc. By mapping all these characteristics into the Personal Awareness software, they can generate automated reactions of the virtual character to its environment in real-time. When the character is presented with several objects or other characters that have been defined as interesting, dangerous, annoying, or whatever, the character will be able to look at them, track their motion in space, and give them maximum or minimum attention. His emotions towards his target will also be reflected on his face and in his body language. Using sliders to define the level and type of reaction of the character, it is easy to create very different personalities that will automatically react to their environment in human-like and truly believable ways. An entire character animation is automatically generated and can be used throughout an entire video game, or even a film for secondary animation. As Frank Vitz says, "Animators will more and more become performance coaches of virtual characters".

Popovic also told about his forthcoming SIGGRAPH presentation on an automated lip-synching tool that will be able to apply any speech audio file on any facial 3D mesh to animate it with lip synch as well as expressive attitudes.