The big trends that will define 2019
It’s that time of year where Foundry gazes into our crystal ball and tries to predict some key trends that could shape VFX (and other areas of tech) over the next 12 months.
1. VR finds its home (out of home)
Throughout last year, we explored the various different ways companies are testing out VR for everything from creating immersive brand experiences to enterprise applications.
The progress of VR to date has generally been measured by the number of VR headsets sold to individuals for next-generation gaming and entertainment at home.
However everything we have seen indicates that VRs most significant impact will be outside the consumer sector (at least in the short term).
Adoption of VR training saw significant uplift in 2018, with everyone from Walmart to Airbus going all in. Similarly, the technology’s use in healthcare has steadily increased.
For those of us in VFX, one area of particular interest that’s on an upward trajectory is location-based entertainment, such as the experiences created by The Void. They’re the VR entertainment company that partnered with ILMxLAB, Lucasfilm’s immersive entertainment division, to create a Star Wars-themed experience last year.
For 2019, location-based entertainment VR is likely to be increasingly tied to film releases, giving studios a way to extend their event movies by allowing fans to engage with the universes they create.
The Void are developing a slate of five new experiences based on material from Walt Disney Animation Studios and Marvel Studios - the first of which has already been released for the Disney animated movie Ralph Breaks the Internet.
2. Deep learning becomes more pervasive
Deep learning works by analysing huge datasets to allow computers to learn useful patterns for problem-solving.
While VR tends to get the headlines, it’s deep learning that is likely to have the bigger and more immediate impact on our day-to-day lives. It already helps us behind the scenes with some of our most common tasks - for things like serving up more accurate web search results, or enabling Siri to understand what you’re asking her.
One source estimates the deep learning market will exceed $18 billion by 2024, growing at a CAGR of 42%. In 2019 therefore, we expect to see a sharp rise in the breadth of application of deep learning.
Some of the areas to watch out for include further advances in self-driving cars, healthcare treatment and illness diagnosis, oil and gas drilling equipment maintenance, and improved protection against viruses and malware - as well as applications within the VFX industry. Speaking of which...
3. Digital humans become even more lifelike
Whether you were impressed or creeped out by ILM’s controversial digital humans in Rogue One featuring the late Peter Cushing’s Grand Moff Tarkin, there’s no denying that photoreal facial animation has come on leaps and bounds in recent years.
In 2019, deep learning (and AI techniques in general) are set to take us even further towards bridging the uncanny valley.
Digital Domain has a whole department dedicated to creating realistic virtual humans - the Digital Human Group. Instead of creating a bunch of facial expressions, putting them into a rig, then animating the rig and seeing what it can do, now they can capture the way the faces moves from an actor from a single camera and train a computer system to automatically recognise, understand and recreate a face in very high resolution, without doing any of the decomposition or compression which can lose the realism of the actor’s performance. And it can be done instantly, rather than in months.
Similarly, companies like Synthesia Technologies are offering ‘professional face replacement’ services: allowing for the seamlessly changing of a presenter or actor's expressions and dialogue. They employ a new method of translating video content that utilises Machine Learning to synchronise the lip movements of an actor to a new dialogue track. This enables translation without the problems of dubbing and mismatching lip sync. The resulting output is ideally completely seamless to the viewer.
Going forward in 2019, we expect digital humans to become ever more life-like - to the point where they're indistinguishable from the real thing.
4. Practical film sets become increasingly sophisticated
Previously we’ve explored how burgeoning virtual and real-time production techniques are being used to create virtual sets. At the moment these techniques are only used by a few pioneering creators - the vast majority of films still use traditional, physical sets.
That’s not to say these sets have not evolved. Under the supervision of ILM’s Rob Bredow, many of the sequences in Solo: A Star Wars Story were filmed using a modern version of the century-old rear projection technique, which harnessed the power of multiple high res 4K laser projectors.
Similarly, the age-old technique of using miniatures to create sets has been given a modern twist, with a period scale model of Brooklyn made for The Greatest Showman. The mini Brooklyn was used for a complex fly-over and built on a large illuminated table, to accentuate the city lights rising up between the buildings.
For 2017’s Ghost in the Shell, eight-foot-models of the film’s dystopian cityscape and exterior environment were 3D printed, enabling the VFX team to plan out the film’s setting before rendering it in CGI and physically walk around the soon-to-be digital setting.
We expect more and more sophisticated techniques like these to be used in the production of films in 2019.
5. All films to be The Jungle Book from now on
In 2016, we had Jon Favreau’s visually spectacular The Jungle Book.
In 2018, the exact same story was re-told in Andy Serkis’ Mowgli.
In 2019, every single film released will be based in the Indian jungle and feature a talking snake.