NBCUni 7.26

An artist’s view of SIGGRAPH 2019

By Andy Brown

While I’ve been lucky enough to visit NAB and IBC several times over the years, this was my first SIGGRAPH. Of course, there are similarities. There are lots of booths, lots of demos, lots of branded T-shirts, lots of pairs of black jeans and a lot of beards. I fit right in. I know we’re not all the same, but we certainly looked like it. (The stats regarding women and diversity in VFX are pretty poor, but that’s another topic.)

Andy Brown

You spend your whole career in one industry and I guess you all start to look more and more like each other. That’s partly the problem for the people selling stuff at SIGGRAPH.

There were plenty of compositing demos from of all sorts of software. (Blackmagic was running a hands-on class for 20 people at a time.) I’m a Flame artist, so I think that Autodesk’s offering is best, obviously. Everyone’s compositing tool can play back large files and color correct, composite, edit, track and deliver, so in the midst of a buzzy trade show, the differences feel far fewer than the similarities.

Mocap
Take the world of tracking and motion capture as another example. There were more booths demonstrating tracking and motion capture than anything in the main hall, and all that tech came in different shapes and sizes and an interesting mix of hardware and software.

The motion capture solution required for a Hollywood movie isn’t the same as the one to create a live avatar on your phone, however. That’s where it gets interesting. There are solutions that can capture and translate the movement of everything from your fingers to your entire body using hardware from an iPhone X to a full 360-camera array. Some solutions used tracking ball markers, some used strips in the bodysuit and some used tiny proximity sensors, but the results were all really impressive.

Vicon

Vicon

Some tracking solution companies had different versions of their software and hardware. If you don’t need all of the cameras and all of the accuracy, then there’s a basic version for you. But if you need everything to be perfectly tracked in real time, then go for the full-on pro version with all the bells and whistles. I had a go at live-animating a monkey using just my hands, and apart from ending with him licking a banana in a highly inappropriate manner, I think it worked pretty well.

AR/VR
AR and VR were everywhere, too. You couldn’t throw a peanut across the room without hitting someone wearing a VR headset. They’d probably be able to bat it away whilst thinking they were Joe Root or Max Muncy (I had to Google him), with the real peanut being replaced with a red or white leather projectile. Haptic feedback made a few appearances, too, so expect to be able to feel those virtual objects very soon. Some of the biggest queues were at the North stand where the company had glasses that looked like the glasses everyone was wearing already (like mine, obviously) except the glasses incorporated a head-up display. I have mixed feelings about this. Google Glass didn’t last very long for a reason, although I don’t think North’s glasses have a camera in them, which makes things feel a bit more comfortable.

Nvidia

Data
One of the central themes for me was data, data and even more data. Whether you are interested in how to capture it, store it, unravel it, play it back or distribute it, there was a stand for you. This mass of data was being managed by really intelligent components and software. I was expecting to be writing all about artificial intelligence and machine learning from the show, and it’s true that there was a lot of software that used machine learning and deep neural networks to create things that looked really cool. Environments created using simple tools looked fabulously realistic because of deep learning. Basic pen strokes could be translated into beautiful pictures because of the power of neural networks. But most of that machine learning is in the background; it’s just doing the work that needs to be done to create the images, lighting and physical reactions that go to make up convincing and realistic images.

The Experience Hall
The Experience Hall was really great because no one was trying to sell me anything. It felt much more like an art gallery than a trade show. There were long waits for some of the exhibits (although not for the golf swing improver that I tried), and it was all really fascinating. I didn’t want to take part in the experiment that recorded your retina scan and made some art out of it, because, well, you know, its my retina scan. I also felt a little reluctant to check out the booth that made light-based animated artwork derived from your date of birth, time of birth and location of birth. But maybe all of these worries are because I’ve just finished watching the Netflix documentary The Great Hack. I can’t help but think that a better source of the data might be something a little less sinister.

The walls of posters back in the main hall described research projects that hadn’t yet made it into full production and gave more insight into what the future might bring. It was all about refinement, creating better algorithms, creating more realistic results. These uses of deep learning and virtual reality were applied to subjects as diverse as translating verbal descriptions into character design, virtual reality therapy for post-stroke patients, relighting portraits and haptic feedback anesthesia training for dental students. The range of the projects was wide. Yet everyone started from the same place, analyzing vast datasets to give more useful results. That brings me back to where I started. We’re all the same, but we’re all different.

Main Image Credit: Mike Tosti


Andy Brown is a Flame artist and creative director of Jogger Studios, a visual effects studio with offices in Los Angeles, New York, San Francisco and London.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.