‘David had the picture of the movie in his head’

The VFX supervisor of ‘Panic Room’–the film is now two decades old–shares key moments from its production.

David Fincher’s Panic Room turns 20 years old this week. The film starring Jodie Foster and Kristen Stewart featured a somewhat memorable troubled production history, partly because the original principal actor Nicole Kidman had to pull out of the project after shooting had began, among other events.

From a visual effects perspective, however, the film is memorable for different reasons. One is the incredible approach taken to extremely long takes inside the main location–a New York brownstone townhouse built on a stage in Redondo Beach–featuring ‘deliberately’ impossible camera moves. These were the result of meticulous previs, motion control and other camera work and a photogrammetry approach to VFX orchestrated by BUF, which had done some similar work on Fincher’s Fight Club.

Another memorable aspect of the film is its unsettling opening titles in which cast and crew names appear as giant lettering framed within New York buildings and locations. The work here was done by Picture Mill and ComputerCafe.

Overseeing those two key visual effects components of Panic Room was visual effects supervisor Kevin Tod Haug, who had also worked with Fincher on Fight Club. He revisits the production in this anniversary chat with befores & afters, looking back at the planning, previs and shoot, and the approach to those impossible camera moves and the unique titles.

b&a: What I remember so vividly about Panic Room was knowing going in that there were a lot of visual effects in the film but that so many people didn’t realize the extent of the VFX after watching it.

Kevin Tod Haug: Yes. You know, a lot of action films used to actively hide the fact that there were visual effects, because they felt somehow it took away from the magic of the stunts and special effects. But that wasn’t the issue with Panic Room, they didn’t even know. They simply did not know. I remember when the FYC ads started showing up in magazines, I remember calling the studio and saying, ‘Aren’t you going to put one in the visual effects rags?’ And they were like, ‘Well, this isn’t a visual effects movie.’ But there were hundreds of shots, which at the time was a big deal. And then if you want to count digital color correction, which was still in its infancy, we had 2000 shots, really.

b&a: Panic Room is of course known for its long shots, its impossible camera moves, which still felt so ‘real’. I remember there was a significant previs effort involved in those. Can you talk a bit about the planning, whether there were boards as well, or something else that David Fincher had done, going into the previs?

Kevin Tod Haug: David did some boards. He does kind of think like a camera, so boards work for him. He had also done a bit of previs near the end of Fight Club. Frankly, Panic Room was the outcome of everything that we had figured out on Fight Club. Remember, Fight Club was, should they be CG refrigerators or should they be macrotures? There were all these tests to try to decide whether or not the CG world was worth touching or not. Panic Room took from that a big use of photogrammetry and texture mapping by BUF. Probably more than anybody would be willing to do now, but at the time, it made a lot of sense.

Then one of the things about planning for that was, in David’s mind, let’s previs the whole thing. David had this plan that he would previs the entire movie. Actually, the first time I talked to David about the film, I was working on The Cell. We were both at the lot, which I think was still Warner Hollywood at the time. I came out to the parking lot and he was like, ‘Hi, how’s it going?’ And then he said, ‘I got some things to talk to you about.’ And I walked over to his car and we stood there for about 40 minutes and he described everything we ultimately did in the movie.

I always say that Directors can be categorised as either Visionaries or Connoisseurs; David had the picture of the movie in his head and he was already trying to download it to people who could get it done for him. This was months and months before the movie was even greenlit. Yet, he was describing shots, like going through the handle of the teapot, through the keyhole…all those things were clearly important right from the start.

He had this plan, in his head, that he could make the movie in less than three months. Of course, it was shot over the course of three calendar years.

b&a: This was because of the change in actors, originally, right?

Kevin Tod Haug: Well, there were several things. Including a change of DPs, losing the Production Designer, the Writers strike, even the set. We built this four-story set of the brownstone where it was always raining outdoors and at some point or other, there was black mold making people sick and that took time to fix. We lost time because we were so damn over. Nobody had done anything wrong, but you just can’t keep something inside wet for months and months and months on end.

And yes, Nicole Kidman was cast, then she had the ‘knee injury’. So we had the force majeure, then we came back and shot for a few weeks, and then she left the production. We were coloring dailies of Nicole and thinking about the film, and still doing previs through all this time. Previs, previs, previs…

I remember sitting in the back of a color-suite while David was doing…something, because I wanted to show him some previs and that morning I had got the rumour that Jodie Foster was on the movie. So I asked him and he said, ‘Yeah…’. His concern at that time was that we had the previs down to the centimeter, literally, so there was talk of putting a box on set that was the exact difference in their heights, so that Jodie could line up to the shots that we’d created.

And then, of course, Jodie was pregnant. She’d said, ‘I’m going to have to get into the sweater at some point or other before this movie’s done.’ And she didn’t in fact make it, we had to shoot final scenes after she’d had the baby. And, you know, Kristen started out this much shorter [shows height] than Jodie and ended up this much taller.

Anyway, for previs, PLF came in and we previs’d for months and months and months.

Previs frame from PLF.

b&a: What were you displaying for the crew, for David, the previs on? Was it a laptop?

Kevin Tod Haug: Yeah, it was just on a laptop. Mostly, we would put something on a USB stick. It was pretty light stuff. There wasn’t yet this idea that previs was somehow a little version of the movie. It was really very much ‘sketches’. It was what a blueprint is to a building, so the actors were all the right height and everything, but they were just ‘Gumbys’. They just stood there like chess pieces, and then they popped into a pose and they slid to the next spot and then they stopped. You could get the eyelines of things right, but they didn’t pretend to act, they didn’t have rigs inside of them. You just changed the pose and off you went to another position.

What it was super-useful for was lining up cameras, and you could be absolutely accurate. In fact, we output a lot of the previs directly to motion control and used it that way. One of my favourite reminiscences was, we were still shooting and one day I’m walking out to set for something that’s coming up and I see our motion control crew walking towards me from stage. I said, ‘What are you doing here?’ Like, why didn’t I know you were here, what’s going on? And they were like, ‘Oh, we wrapped.’

I said, ‘So what were you shooting?’ And they said, ‘Well, David was just using motion control for something.’ So, the thing was, David sometimes preferred motion control to operators because they could repeat the exact same move this way. Once he got it right, he didn’t have to worry about it again, it just repeated itself. And for a guy who is known for 20, or on this movie, as many as 50 takes, having one of the parameters always be the same, that meant he could just focus on the other thing, the acting. It was a big deal for him and it was pretty quiet that rig, so it wasn’t horrible down stream.

b&a: How has previs changed since you did it that way, do you think?

Kevin Tod Haug: The big difference is that previs was always a victim to the Analog Abyss (aka production). You would do all this great stuff in prep, but because there was no way to keep it and pass it through to post, it would die when you went to set, which is one of the promises of the new, Real Time, paradigm.

With the film I’m working on now, Comandante, the goal is that all of the previs informs and becomes at least the background. It’s not something where you kind of eyeball it or take some information. It’s actually those shots, those assets have evolved from the very beginning to what you have on ‘The Screen’. In 2000 there was not enough resolution in previs to do that and since you knew it was just informative and therefore disposable, you didn’t do anything more than you actually had to. Except those motion-control moves; they ended up on The Screen. I have been chasing that paradigm ever since.

Techvis setup for motion control shot, from PLF.

b&a: What was the state of play then, in terms of on-set stuff that needed to be done, in terms of on-set surveying, LEDs or markers?

Kevin Tod Haug: We did try LEDs for tracking markers. The LEDs we were using had been used on Batman & Robin. They used to be attached to Arnold Schwarzenegger when he was Mr. Freeze, and they were red and they were incredibly bright, and there was no way to dim them. And so they actually flared the lens and at some point we had to eject that whole plan, because there was no way to tone them down. We were wrapping little gels around them to try to bring down that flaring.

Tracking at the time was still kind of a 2D operation. BUF, I think, used the whole bag of tricks for the ‘Big Shot’, as we called it, the one where we go all the way through the house. Mostly it was tape on the walls for tracking that just got removed later.

b&a: What were BUF’s requirements, in terms of their photogrammetry, on set?

Kevin Tod Haug: Well, we covered that house to death with stills photography. Now, there had been a couple of issues with the ‘sport fucking’ scene in Fight Club. BUF had always used very low ASA stills film. The problem with that is that there tends to be about three stops of dynamic range on stills film and the DP was used to working with the normal stock that he’s used to working with, which is not nearly as fine grained, but it’s upwards of seven stops of ‘latitude’. So, some of the stuff was incredibly underexposed. This was what Jeff, (Cronenweth, the DOP on Fight Club) had to work with. He had to light a scene and the photogrammetry work needed to fit into it. He did his best to match them but he couldn’t realistically relight the movie for the stills film. So we made some compromises, and some things had to be completely remade in CG.

Wireframe view of the ‘through the coffee mug’ shot by BUF. For several years, the VFX studio had pioneered stereography and photogrammetry methods for reconstructing virtual and, importantly, photoreal sets, enabling the camera to travel anywhere in the scene.

The final shot.

We didn’t want to do that on Panic Room. Knowing this, I went to Imagica where Doug Trumbull was scanning 65mm back in the day. Now, when you bought film in those days, at some point you commit and you go out and you buy your stock and you buy all of it from the same emulsion batch, right? It’s the same as LEDs now, you have all of it from one batch. If you get another batch, it’s not going to look the same. I ordered a couple of cans of 65 from the same batch as Darius (Khondj, original DOP of Panic Room) ordered for the movie; which we then put into an IMAX animation camera that Les Bernstien jury-rigged into a stills camera.

The reason we were doing 65 was because of the grain issue that BUF had solved by using finer-grained still film. We were using the same stock as Darius Khondji, and later Conrad Hall were shooting with, exactly the same batch of emulsion, but 65mm which, effectively, has no grain.

Sooner or later, every room was lit, so we had this catalog and we would go in and cover the room with stills photography every time it was lit. For every setup, we would go back in and try to get some of it before they moved on. This meant we had a double lighting kit just because of us. It was a deal I made with Darius that I think ended up being part of the reason we had three setups all the time, because there were all these extra lights laying around that could be used.

They’d finish the setup, and then they would move and we would stay, and they couldn’t take the lights with them because we needed to stay and shoot our part. The extravagance of this kind of filmmaking sometimes boggles my mind, but we did do all of that and it worked pretty well. In the end, we used texture maps on everything. A lot of the stuff is not shaders, instead, a lot of the stuff you’re looking at is literally texture maps.

b&a: I wanted to ask you about the opening titles, too. Did you have direct involvement in that, because that’s not always the visual effects supervisors purview?

Kevin Tod Haug: Here, it was, because it was a Fincher movie. I also came from graphics, into TV, back through graphics and eventually into movies. Picture Mill was the title design house, and ComputerCafe was the VFX house on this, and then Bill Lebeda from Picture Mill was creative director.

It’s interesting, there are legal issues with making type different sizes on title sequences. So we came up with a legal thing where the actors had to say, it’s okay, as long as it’s on average or scale-wise equal to each other. But scale to what? Well, scale to the building it’s on. Then we were safe. It was tricky legally, to do that.

b&a: Were these titles boarded and mocked up?

Kevin Tod Haug: Yes. The idea for them was the first thing that Fincher ever told me, long before talking about the big shots. He said, ‘I want to do the titles as if they were a Christo installation. I want them big, floating in the city, in place and look absolutely like they belong. They’ll just be white. I want the color from the building across the way to shine on them and all that stuff.’

Ultimately, I shot those plates without exactly knowing which was going to go where, and they were shot near the end of the movie. When Jodie was out on maternity leave I went with a second unit and ran around Manhattan and shot all those plates.

b&a: Was there anything you had to do to get those plates to feel a little like aerials. They’re not aerials, are they, but they feel like it?

Kevin Tod Haug: No, they’re building to building. They’re absolutely locked off plates building to building. That’s the way Fincher wanted it to be. He wanted it to start up high downtown and work its way slowly towards the park. And then eventually we’d end up at the brownstones. He’d picked some of these places specifically, and the other ones we found because they filled the bill. For a while, my credit, which never got on the movie, was attached to a building in Gramercy Park. It’s actually funny because when I worked on Stay, I stayed in the building that we used to shoot what would’ve been my credit. I could see where my credit would’ve been, out my window.

b&a: Those credits, of course, just felt so unsettling and I’ve never quite understood why, it’s the music as well, but they feel so unsettling. It’s such a great opening.

Kevin Tod Haug: David wanted it to feel like Vertigo. He wanted it to feel like his version of those kinds of movies, which I actually think is a really good place for Fincher to go. He kind of naturally lives there and he can do that kind of thing like nobody else can. It does give you this strange POV thing. You know, ‘Who’s is the POV?’ as we’re moving downtown. And then the POV goes in the house and occasionally the POV flies around the house. It’s like, who is the ghost? It’s us.