Search form

No Time to Chill: Sony Imageworks Cranks It Up for ‘Ghostbusters: Frozen Empire’

VFX Supervisor Jason Greenblum talks the singular challenges and dedicated craftsmanship involved in the creation of the visual effects for the latest iteration of the beloved Sony franchise.

It’s truly a cold day in hell when, after an ancient artifact unleashes an evil force, an army of ghosts casts a death chill upon New York City in Sony Pictures’ Ghostbusters: Frozen Empire. Who could have foreseen that anything would go wrong when the Spengler family teams up with the original Ghostbusters, who’ve developed a top-secret research lab to take busting ghosts to the next level? But now, here they all are, joining forces to protect their home and save the world from a second Ice Age.

Featuring a glittering (you know, like ice) cast that includes original Ghostbusters Bill Murray and Dan Aykroyd, with direction by Gil Kenan and a screenplay by Kenan and Jason Reitman, Frozen Empire is a mélange of old and new, with enough spooks, gags, and wacky mayhem to keep things running at a lively clip. But all those battles and icy objects don’t create themselves. Accordingly, we spoke with Sony Pictures Imageworks VFX Supervisor Jason Greenblum about exactly how they did get created, which, as it turns out, also involved a spirited interplay between the present and the past.

But first, enjoy the trailer:

Dan Sarto: So, in the new film, we’ve got lots of ice, lots of things that get frozen and then shattered. New ghosts, familiar haunts, little Mini-Pufts that go back to the original giant marshmallow. Can you give us an overview of the main visual effects that you did on the film?

Jason Greenblum: Sure. This was a unique case for us because we actually did the majority of the shots on the film. There were about 1,300, I think, and originally we were going to take the whole thing on ourselves. But time got crunched, and so we had to find some chunks that could be parceled out to other facilities. We ended up doing about 1,070 shots. We did everything from split screens to all the Mini-Pufts running around, Garakka – the demonic ice deity, and Patience the Lion. So, we had massive effects shots with storms – the storm coming in, the summer storm – and all the effects associated with Garakka.

It was fun for us, because, as Imageworks, we hadn't done a full show in a long time. So, we had to break out some old techniques and rethink some new paradigms. It's easier when you have one environment, a couple of characters, and you can build pipelines around that and then crank out the work. But, with this, it was so varied. For example, for the Sewer Dragon sequence, we did a bunch of car comps, which we hadn't done in a while – so it was refreshing in that respect. We got to sharpen our pencil on pretty much every VFX trope there is.

DS: I'm assuming you guys came on at the very beginning. How much testing did you do ahead of time? Did you do previs, or did you have previs to work with? Did you have to do any of your own designs? How did the process play out?

JG: With certain things, we had some concept work done. Obviously, they had done a bunch of pre-production. For instance, for Garakka, there was actually a sculpt and 2D illustrations. And we had to take it out of that two-or-three-still-image world and bring it into a three-dimensional world. You're looking at the character in a loose sketch, and then you have to realize it up close and tight, and at different angles… how the mouth is going to open, and how the eyes are going to move. So, we had to go through all of that on our own side.

We did involve some concept-painter people in fleshing out the face and the proportions. And then things evolve, obviously. Originally, Garakka was meant to be much more vaporous at the beginning and hardly physical, and he was going to change through the course of the movie into a physical being. But the way it played out, he ended up coming into the world more physical than expected. So, there were more scenes in which his skirt, torso, and skin were prominent features.

For the frozen city shots, we just basically conceptualized those in-shot. And a lot of the effects that we did for that final battle with Garakka, the family, and the Ghostbusters – the different aspects of Garakka’s freeze breath, or the death chill, the tri-stream, the modified proton streams from Phoebe's wand – that all came together in-shot. There were some characters, like Pukey, that were basically designed – we had some artwork and we just made it three-dimensional and brought it to life. So, it ran the gamut.


DS: What about previs?

JG: No, we didn’t do the previs. I think DNEG did some early previs for some of the scenes, and then The Third Floor came in and did a bunch of previs for the postvis kind of stuff.

DS: How much of the New York environs was actually shot and how much was recreated digitally?

JG: That's a good one. For the car chase with the Sewer Dragon, I would say predominantly that was shot in New York, with an array car, and the Ecto-1 running through streets. We did have to recreate a lot of the stuff. Inside the car, when they're driving around, was a system built internally for specifically doing that work with the array car that they shot with.

On the other hand, the library was almost entirely synthetic. Unfortunately, due to timing and scheduling issues, they had to shoot all of that on a partially built set outside of London, in Reading, and it’s basically a parking lot. It probably occupied one-tenth of the space that it needed to take up to do this work. So, the library itself was entirely CG. That was our creation. The trees next to it, down the streets, left and right.

And then, when you looked out from the library across the way, that was a combination of CG enhancements and array footage that they shot in New York, separately. It was a three-camera array that was trying to capture moving footage of cars driving by, and people walking through, and the life of Manhattan. So that section was a real challenge, just getting it all working together well.

And then, for in front of the firehouse, we digitally recreated the entire four city blocks there, just because of the nature of the work we were doing, where we were freezing it over, and the fact that we had limited access to it during the shoot. There were some plates shot, some reference material; you had to enhance or augment the facade of the firehouse because the actual practical firehouse in New York has different dressing and then paint on the street and stuff. But that whole area in front of the firehouse, when you see it in the movie, is CG.

DS: The production notes mention that the filmmakers wanted this movie to tie into the visual style of previous films, even going back to the original. How much were you able to do that? And how difficult was it?

JG: That's a great question. I got involved with this early on, back in January of last year, I think. And I had a call with Gil Kenan, the director, who was very specific about his desire for this film to have that feeling of the 1984 movie, the feeling that the effects could have been done practically. He really wanted us to look at everything we did through the lens of what they would have done on set practically in '84.

Obviously, technology's advanced and we have different tools now, so we wouldn't do it that way. But the idea was to keep the texture of the original movie. So, for example, in the case of Garakka, they would've built him as a prosthetic, which would've been painted silicone. So, when we created Garakka, we wanted it to look like this was something that could have been photographed, something that was mechanical, but we also wanted him to feel alive and scary and real.

So, it was very much through the lens of how would they have done it then. What can we take from that – even to the point where, with Melody, we could now do so much with 3D tracking. We did end up tracking her, but we wanted it to feel like it could have been optically printed. Gil was very specific. He didn't want to have the dimensionality that you could get from a 3D object and tracing the backside of a surface. Procedural looks were a no-no. He wanted it to feel physical and of that era. So, we actually used ideas derived from optically printing something – how would that look, and what parts would you lose, and what parts would you keep? And it was very much driven by that.

DS: What about practical effects – puppets and things like that? Were they used at all?

JG: That's another excellent question. Along those lines, Slimer is a perfect example. They went back to the original molds, and they had a new casting of Slimer the puppet built. There was a point at which [writer and son of original Ghostbusters director Ivan Reitman] Jason Reitman, who of course grew up with the Ghostbusters, was almost brought to tears looking at this puppet on set. And they had these great puppeteers working on it. So, we shot bluescreen elements of Slimer for every shot, and we used them with the actors, in order to capture really authentic responses.

Ultimately, because of what we were trying to achieve with the shots, and the limitations of a puppeteer in a suit, and the spaces that we were dealing with, we ended up augmenting or replacing synthetically pretty much all the Slimer shots. But we absolutely used the photography to inform how he would animate, how he moved, the limitations of his movement – it was excellent material to draw from. And, a lot of times, it's simply RotoMated. We didn't enhance the animation much, we just cleaned it up.

DS: Among the many effects you did on the film, are there any sequences that stand out as being the most challenging? Or sequences that wound up being more difficult than you anticipated – or ended up being easier that you expected them to be?

JG: The final battle between Garakka and the Ghostbusters was especially challenging, mostly because it had to come together very quickly. They had some reshoots that really filled out the sequence, and there were some changes that had to get plugged in, in order for us to move forward with that sequence. So, we ended up doing that in a fairly short period of time.

And there was quite a bit of effects work involved with Garakka's capture, with the fire-bending from Nadine, the tri-streams from Phoebe's modified proton wand, all the ghosts trapped that came out of the containment unit. Almost every shot had effects in it, and a lot of it had to be created in-shot and figured out in-shot. So that sequence probably offered the most challenges.

However, there were some scenes that I thought were going to be quite difficult that came together much more quickly than I expected. One example is the large, ominous cloud that Garakka brings with him, that overtakes New York in the summer storm sequence, where it kind of flows in. That could have been very challenging, but, again, we went with ideas from the original film, where they probably would have shot some of this stuff in a water tank and just used water tank footage, and slowed it down, and sped it up, and flashed lightning in there and stuff. So, it came together a lot faster than I expected, and I thought it looked great.

There were so many effects that we had to come up with in-shot. Some of the stuff in the lab, when the Ionic Separator has to extract ghosts out of various objects – that was really just fleshed out in the shot, and those came together really well, I thought. We had some stills for that from pre-pro, but, in motion, you have to really figure those things out again.

DS: By now, I imagine you guys have done basically everything, but that doesn't mean you necessarily have done it recently or are set to do it. Did you have to come up with new tech, or modify your pipeline or your tool set, for this show?

JG: That's another good question. One of the things that actually was more difficult than I thought it would be was the ice spear work, which actually played a huge role in the movie. And you might think, "Oh, it's just spikes growing out of the ground." But it presented a lot of challenges we didn't expect. It was really complicated and tricky to get the perspective right, and the sense of the spheres coming at camera or away from camera. Gil really wanted to convey the idea that those spheres were following Garakka and always had a very clear directionality. And it was a lot more challenging than anyone expected to get that sense of them flowing in a direction, and the scale, and the movement right.

So that was one of the things we had to build a pipeline around, and we had to build special tech for how those got rendered, and how the textures were scaled up and were kept consistent. We worked quite hard on getting those ice spheres to have the right visual quality, especially because there was also stuff that was photographed on set that they had to match. They were plastic cones with some cellophane or something inside to give them that volume. So, we had to match that or at least create something not too dissimilar from that.

An animation would set the timing and the scale of things, and that iterated several times through. And then effects took them over and would create extra debris and rubble around them, and create bursts when they punched through things, and lighting had a big role in that look. But to look at it beforehand, you wouldn’t have thought it would be challenging.

Also, while it wasn’t necessarily new or bleeding edge, I thought the Garakka secondary character performance stuff that we did looked really good, and subtly maintained a sense of realism. He's this skinny, scrawny creature, but he also has muscles and he’s really powerful. We had a whole team that used Ziva or Sculpt to create the sense of flexing and the skin movement over the surfaces.

DS: If you had to summarize the experience of working on this show, what would you say?

JG: I'm really proud of the work that Imageworks did, especially considering how varied the work was and that a lot of it had to be created in-shot without clear direction. I think it's a testament to the Imageworks team's creativity, everyone from the artists all the way up to supervision. Because of the amount of work we had to push through, we broke into two separate teams, and we had two flows, and two effects supervisors.

Spencer Lueders and Chris Messineo were the effects supervisors, and they had to funnel through a lot of work, and create interesting looks, different looks for unique, varying shots. There were so many one-offs. I think that’s what really impressed me on this show… the level of creativity, including the lighting and compositing teams. Everyone really contributed, everyone cared about their shots, and it really showed in the work.

Dan Sarto's picture

Dan Sarto is Publisher and Editor-in-Chief of Animation World Network.