The film Limitless topped the box office after its opening week, and the movie's intro sequence may go down in history as one of the most memorable and immersive shots of all time.
Limitless opens with a continuous, seemingly endless zoom shot that starts as if the camera had been dropped off a skyscraper balcony, then weaves through the streets of New York, past pedestrians, under construction scaffoldings, though blocks and blocks of blinking lights and hubbub, through the back windows of taxicabs and out through their front windshields, through Harlem, into Times Square, and then right into a glowing Jumbotron. You are gliding at street level right through solid objects, and the shot reveals little visual evidence of computer-generated trickery.
But the shot doesn't end there. It keeps going, sharply and steadily through the inner workings of a human brain, which then morphs into a bird's-eye view of Manhattan, drops onto the roof of a high-rise, and somehow leaves you looking up at the sky. This is not the kind of sequence that someone can shoot with a regular camera. In fact, this is not the kind of sequence that even the Hubble telescope could shoot--unless it was somehow made of a subatomic particle, had flawless continuous autofocus, packed a bajillion-megapixel sensor, and also had teleportation powers.
This is the kind of sequence that makes you wonder, "How did they do that?"
After seeing the film, I was determined to find out. I spoke to Comen VFX visual effects producer Josh Comen and visual effects supervisor Tim Carras, both of whom worked on the title sequence and other effects as part of the Limitless crew. Comen and Carras led the team that stitched together, rotoscoped, and composited the sequence, and they provided some fascinating insight on how it came together, from the script to the direction to the camera setup to the extensive post-production process.
PCW: How did the idea for the sequence come about?
Tim Carras: The whole idea for the sequence came from Neil [Burger], the director of the film. His term for it was "fractal zoom." If you've ever seen those Mandelbrot fractal patterns--[they're] like blobby, amorphous shapes, but as you zoom in you see that the small details are the same shape as the objects that they came from. So you can keep continuing infinitely into the details, and they keep growing into the bigger shape that you saw before.
The concept was something he'd been thinking about for years, and with this film he finally had a chance to put it into practice. He's the one who came up with the concept and the look of the shot.
PCW: Was that something that was in the script, or was it sort of added after the fact?
Carras: It was something Neil added to the shooting script. It basically says, "Camera drops down into New York street, and through a series of fractal zooms, we see X, Y, and Z." He wanted a visual-effects overture to the film, and he knew the title sequence was the place to do it.
Seeing a descriptive line like that in a script is absolutely the fun part of our job. Starting out on the conversation of how it's going to be realized visually is a tremendous thrill, because that's the creative essence of the work we do: taking words on a page and making them into something an audience can watch. Hopefully, they've never seen something like that before.
PCW: Do you want to just dive right in and tell the story of how the sequence was created?
Josh Comen: I think if you only use one technique to accomplish a sequence like this, the audience is going to get it, because the audience is always more tuned in than one assumes they are.
Carras: Here's the big secret. Everything you see in that infinite zoom in that scene through New York was shot on completely stationary cameras.
The way it was shot is that there was a rig of three Red cameras mounted side by side on a single tripod, and each one had a different lens on it ... a wide-angle section of the street, a medium, and a close-up. All at the same time, [they're recording] the same movement from the extras walking around in the frame and the cars traveling down the street. And because it's New York, you've also got the lights and the billboards blinking on and off. Just tons and tons of motion in the frame. So it's important to capture all that video with three lenses at once so you have all that information.
They did the three-camera setup at every block and intersection in several parts of town. The title sequence opens on 8th Avenue, then it goes through Harlem, and through Times Square.
PCW: So is the sequence geographically accurate from start to finish?
Carras: (Laughs) Not in the least. You couldn't actually follow a straight line from 8th Avenue to Harlem ... but hopefully that's sort of left out of the equation when you're watching it.
If you work out the mathematics of it, which we had to do inside our software to get it to work, it's essentially a zoom lens that goes from a 25mm to a 10-to-the-36th-power millimeter.
PCW: (Laughs) Those lenses aren't readily available.
Carras: (Laughs) Not in the real world, no.
There was a lot of discussion about whether there would be motion blur as the camera is moving in, because in real life, if you're doing a zoom that fast, everything would kind of start to streak out and blur. But Neil preferred that we do it with no blur at all, so that it was all very crisp and sharp and pristine, and also because I think he had the idea that we could use that motion blur to hide transitions.
(Laughs) He didn't want to give us any excuse to hide anything. That was a good call, because as Josh said, audiences start to get wise about the tricks that they've seen before. If it was blurry, people might have thought, "Oh, that's so they can hide all their mistakes."
PCW: How many individual camera setups were used for the sequence?
Carras: Not counting the cabs, which were their own separate thing, I think there were probably about 15 camera setups. Times three angles per setup, so 45 different camera perspectives that we used.
And that's just the first third of the sequence. Once we go through the Jumbotron in Times Square, the Jumbotron ad turns into an MRI scan of a brain. That was a whole different sequence that we designed in-house. We had a team of animators building brains and neurons, and finding objects we could use to create textures that look like the inside of a brain.
PCW: And then the brain morphs into a bird's-eye view of Manhattan?
Carras: Yeah, the last part is the entire greater Manhattan area from overhead, and then we go in through the rooftop of a building, and that turns into the sky. Pretty crazy sequence.
They also had a single Red camera in the backseat of a cab driving around New York as well, which came in handy.
PCW: For the parts of the sequence that went through the back of the cab and out the front window?
Carras: Yup. I'm still not exactly sure how that came about. Before we'd even started putting anything as a visual effect, we were just combing through the footage and stringing together a rough animatic of the structure of the sequence, sort of what the bones are. And Daniel Durand, our in-house visual-effects editor, pulled all these shots from inside the cab, and we said to ourselves, "Wouldn't it be cool if we traveled through all these cabs."
PCW: How did you make the transitions between those different locations look so seamless and continuous?
Carras: The transition between 8th Avenue and Harlem ended up being a string of six taxis and one cop car and one civilian sedan... Those were done a bit differently, because we didn't have our three-camera setup looking at the back of a bunch of parked cabs in a row. So those were all found footage from the single camera in the backseat of a cab driving around town. Any time the cab came to a stop at an intersection with another cab in front of it, we grabbed the frame of the backseat of the cab with the driver looking in the rear-view mirror and the taxi meter and the fiberglass cage. We took a whole bunch of those and stacked them up so it looked like you were traveling through the windows of a sequence of cabs, one after another.
If you watch the sequence really closely, it's always the same cab driver in all the cabs. You can see the guy's eyes in the rear-view mirror. That's one of our giveaways. It's always the same guy.
And actually, the cop car is a cab that we changed the colors of. We went in and researched what New York City cop cars looked like and what font the word "Police" is in on the back of it, and the bumper stickers they have on the back.
Comen: We didn't want to have the typical "D.A.R.E." sticker, because I actually called the police department and looked into it, and they don't use those stickers on their police cars. They have a bumper sticker that's specific to the NYPD, so we used it. It adds to the sense of authenticity, and the last thing we want to do is make any alteration to the film that isn't authentic.
PCW: Was it hard to make the body of the cab look like a cop car?
Carras: The idea of changing the cab to a cop car came from Josh, because he knows that most cabs are actually former cop cars. Once they sell them, they already have the cage separating the front from the backseat, so it's easier to remodel them into cabs.
Comen: If you look at some cabs, they say "Police Interceptor" on the back--obviously just a cop car that's out of service, so it was the path of least resistance for us that worked for the production. And it was ultimately what they wanted, so it worked all the way around.
Carras: One of the early notes we got was, "Does there have to be so many cabs in a row? Can we change some of them?" So we said, "Sure, we'll change one of them to a cop car." And we managed to find an Accord in the footage somewhere, and it happened to be stationary enough to work.
PCW: What kind of software did your team use for the effects?
Comen: The primary software we used was [The Foundry] Nuke for the compositing, [Adobe] After Effects for the titles, and Maya for the 3D graphics. Our team was a crew of about 20 people, including support staff.
PCW: How long did it take the team to produce the whole sequence?
Carras: We started right after the holiday break, so that was January 5, and we ended delivering the final in the first week of [March].
Comen: And we started talking about it in December. We started conceptualizing when we were reviewing the FX shots with Neil as early as late November or early December.
PCW: As a visual-effects team, did your role overlap with the director of photography and the camera crew for a sequence like this?
Comen: No--we came on board after principal photography was complete. The main role here that I had was making sure our entire staff was constantly working, and basically available 16 out of every 24 hours. Any time production made a change, I couldn't tell them, "Well, no," so we had to roll up our sleeves and make it happen. For a sequence like this, every change has a domino effect. It's such an intricate sequence, you can't change one piece of the sequence without it affecting every other part of the sequence.
Carras: We had one team that was devoted to the MRIs-and-neurons portion of the sequence, and we had another team that was dealing with the satellite imagery, and then a whole different team of compositors that put together all the Red footage of the New York City streets. In terms of sheer amount of manpower, the city streets required the most amount of work, because we were stitching together so many different images and traveling such a great distance. We had to find a lot of organic edges to blend in together with all that live-action footage, and dealing with cars driving and people crossing, and all that had to be rotoscoped frame by frame to make the edges line up perfectly.
PCW: The end result is pretty incredible, and it totally pulls you into the movie from the beginning. You're watching the shot, knowing that it's physically impossible, but everything looks so sharp and realistic.
Comen: "Seamless" has always been our goal, and when we're interviewed about the work we do, people often say, "But wait, I didn't see any visual effects there." And that's when we sort of pat ourselves on the back, because that's what we're aiming for. In Limitless, unlike other feature films that we've done, there are also obvious visual effects. So there was a blend of visual effects being utilized in both seamless and obvious ways, both in the name of moving the plot and story further ahead.
Carras: As you say, it's really hard to tell whether it was done in-camera, digitally, in post-processing, et cetera. It's interesting and rewarding for us to hear a viewer's point of view on the effect.
Comen: When people have no idea how you did it, that's what we want. It's rewarding when someone has no idea how you did it. When the viewer asks, "What was the one thing they did to do this whole thing?" the answer is always that it wasn't just one thing. It's a blend of several different techniques, and that's what ultimately makes the work look good.
Like this? You might also enjoy...
Profiles in Geekdom: The Designer Behind Eightbit.me Move Over, Dr. Soong: Girls Can Build Android (Apps) Too Keyboard Cat Wrist Rest Is Meme-tastic!
Get your GeekTech on: Twitter - Facebook - RSS | Tip us off
Source:
No comments:
Post a Comment