What was actually involved was a massive amount of trial and error, logistical preparation, about 1000 cars arranged in a carefully crafted grid and headlights hooked up to remote transmitters. Van Wyck and his team had about eight weeks to figure it all out and the film was shot over five nights.
We talked to van Wyk about what attracted him to the job, why he decided to do the whole thing in-camera and more.
What were some of your first thoughts when you saw the brief?
It was always about creating this animation using car head lights. So the immediate thinking was whether to do it in-camera or to use post, but then you realize that when you want to work this way with light you can't really achieve your goal with post. You have to do it in-camera. There was so much more happening that we couldn't foresee, so we had to do a lot of tests to see what it would look like. Once we did that we were totally convinced that we were doing it the right way and doing the whole lot in-camera was the way to go.
I knew I wanted to do it in-camera because we've become so post-wise that we're not impressed by it anymore. When you see an example of large scale anything, you immediate instinct is to say they used post, so I think people want stuff that's more real and that was really important to me. Now, it's all well and good saying that but no one knew what to expect so we had to run a huge number of tests. The agency really liked the direction so once that was signed off on it was just finding out how it would all work – the height of the camera crane, the distance between cars, everything. We had to do a lot of CG grids and other kinds of leg work.
Was this like anything you'd worked on before?
No, I'd never done anything like this. It was weird, I got it based on the treatment and I think they liked the look of some of the things on my reel and there was a human element that was really important to me in my treatment. I absolutely loved doing this. There aren't many big scripts like this around, that are this visual. So when this came around I jumped on it with all claws. You don't expect to see another script like this for a while.
What were some of the biggest directorial challenges?
Just doing the math beforehand. Before we even put the grid on the ground, it was just about figuring out what we could see from that height with what lens. We just had an endless amount of models, had to find the right location with the mountains in the background, with the sun coming up over the right side. Because the crane is slightly lower, to get that perfect LED grid, the cars at the back have to be another half kilometer further away. So we ended up dealing with kilometers. And if the cars are too close, the crane would have to be 80 or 90 meters in the air as opposed to 60 meters, which is totally impractical because the wind bounces the crane around. So we kept going up and down trying to find the sweet spots.
We had three different cameras going, all at different exposures, to get the sunset. But I was really happy because once in the edit, it was there with the intensity we wanted. So in post, all we did was play with telecine a little bit.
How were the lights operated?
Most of it was automated. We had to write a program for that. But the car lights respond a certain way. If the program's too fast, the lights don't respond properly then the animation looks blurred. So we had to find a way to play with the speed in-camera. But there wasn't really one formula to achieve it, it was a bit of everything.
In the "Making Of" video, you said the shoot got off to a great start, no doubt due to meticulous preparation. What other parts of the job surprised you?
When you're dealing with so many lights, just the physicality of so many lights, and guys are talking about 10 to 20 kilometers of cable to make it work and we had to go with transmitters that send signals. And with transmitters, it seems anything can happen. So the biggest surprise for me was when we did the test with the entire grid and the camera in the sky, it all came out clear and everything worked, save for a few little things. It was a huge rush to see it. It was incredibly calming to see it all work and not be panicking about all the work there would still be to do in post. To actually see 80 percent of it happen right there in front of you right away was amazing.
It was an incredible project to be a part of. You can really get stuck in it. I'm very happy with how it turned out.