Exploring the state-of-art in how AI is transforming the visual effects (VFX) industry, by Martin Anderson
New machine learning techniques being pioneered at the major visual effects studios promise to transform the visual effects industry in a way not seen since the CGI revolution.
It’s over twenty five years since the ground-breaking CGI effects ofJurassic Parkusurped 100 years of visual effects tradition. When Steven Spielberg showed the first rushes of computer-generated dinosaurs to acclaimed traditional stop-motion animator Phil Tippett (who had been hired to create the dinosaurs in the same way they had been done since the 1920s) he announced“I think I’m extinct.” It’s a line so significant that it made it into the movie itself, in reference to a paleontologist envisaging a world where no-one would need him to theorize about dinosaurs any longer.
Though a quantum leap, the visual effects of Jurassic Park did not represent an overnight upheaval. They had been presaged sporadically throughout the 1970s, and at greater length in the 1980s, in cinematic curios such as Tron, The Last Starfighterand Flight of The Navigator. In the few years directly prior, James Cameron had brought renewed interest to the possibilities of CGI with the ‘liquid’ effects of The Abyss and Terminator 2: Judgement Day.
But Jurassic Park was different: computers had achieved the ability to generate solid,photo-realobjects, promising to relegate the uncomfortable burdens of the physical, photochemical VFX world. It set the trend for the decades ahead, and reinvented the visual effects industry — not without many casualties among the old guard.
Many influential movie makers and VFX studios were unable or unwilling to read the signs of the times in the years leading up to Jurassic Park. It now seems that the water is rippling again for the current state of the art in visual effects, as new machine learning techniques slowly encroach on the now-established workflows of CGI. And that a new ‘disruptive event’ may be coming to shake up the industry.
By Martin Anderson