Mark van den Bergen was the on-set VFX Supervisor on the sci-fi film 2067, directed by Seth Larney. The film tells the story of the year of Ethan Whyte, a tunnel worker for a city’s power plant when the Earth has been devastated by climate change. With all plant life on Earth extinct, civilization has collapsed around the globe; only one city has been barely able to hold out against this catastrophic environment, thanks to synthetic oxygen.
One day, during a test of the Chronicle, a prototype time machine, the scientists receive a radio signal from 407 years in the future with a message to specifically send Ethan to them to prevent the extinction of mankind.
Van den Bergen has worked in VFX features since 1997, including films such as Mad Max Fury Road, as part of the 15 person post-visualisation team under lead VFX sup. Andrew Jackson.
FXG: How did you end up working on the film 2067?
It was good timing; I had just finished up VFX Supervising on the short film Jasper when John Deitz from Bang Bang pictures contacted me to work on an Australian sci-fi film. I came onboard during pre-production and was involved in location scouting then onset VFX supervision followed by post visualization and then I took some of the more tricky shots to final. When I arrived there was already a lot of reference material from the director Seth and Production Designer Jacinta Leong. This included storyboards, mood boards, concept art, and reference materials of other films.
The first meetings were about discussing and evaluating the various ways that we could approach VFX intensive sequences, trying to work out how much of that would fall on VFX and how much would fall on the other departments and how much would be achieved by shooting at specific locations. We did a fair bit of location scouting in the forests of Elands in New South Wales (Australia), where the director grew up as a child. However, those areas were leech and tick-infested, so we spent the evenings picking ticks and leeches off each other, and ended up shooting in places where we could crawl around other forests without blood-sucking bugs.
FXG: The director, Seth Larney, is a very VFX savvy director, how did you find that relationship?
On 2067 we had the advantage of having a writer-director who has also been a VFX Supervisor and VFX artist in the past. Seth has been working on VFX films back to the original Matrix movie, so he really understands the process inside out. When you have a limited budget, there is a lot more creative problem solving to be done because throwing money at a problem is not an option. Seth always had a plan A, a plan B, and a plan C up his sleeve. I can not imagine getting a VFX intensive film like this done on a limited budget without a director that can write scenes with the VFX in mind.
I was in a room at Spectrum studios working on quick and rough VFX shots to feed the edit when the director was working down the corridor. This workflow is flexible enough to keep up with the director’s creative process. The director might wake up one morning with a new idea for a VFX solution and needs to see it the same morning. Having me down the corridor, available to switch tasks by simply having a chat with the director, is a very flexible way of quickly testing VFX solutions in the edit. For example, when the story required some look development to bring some more life to the time machine, I was able to start providing examples immediately and turn those 30 shots over in a couple of days. The postvis workflow means that the director can test visual ideas that are important for the storytelling, to quickly see if they are going to work.
FXG: How did you work with the other departments?
Because the visual effects are done after principal photography has wrapped, there is a danger of the VFX being ‘out of sight, out of mind’. The teams on-set often think of VFX as ‘something that another team will deal with later’.
Part of my job onset is to bring the awareness that even a simple VFX fix can be $1000 or $2000 simply by the fact that the shot has to go through the VFX pipeline. The fact that something is easy to fix in VFX does not mean that it is cheap to fix, especially if there are many shots needing the same ‘simple fix’. Solving things on set, before they become a VFX problem, usually involves making requests to the camera department or art department in pre-production.
Our amazing camera department led by DOP Earle Dresner and Art department led by Production Design Jacinta Leong and Art Director Gareth Wilkes, were an endless source of creative solutions on this film.
It is incredibly important to have a representative from VFX on set and in pre-production to flag those problems to avoid very expensive problems later.
FXG: Was there much green screen used, was this a big comp or 3D project?
We had several scenes with a green screen outside a window where VFX would be putting a dystopian city in the background. There are some green screen shots where showing this dystopian city as an essential part of the storytelling, and other shots where there is a risk that the green screen might just pop into the edge of frame without adding anything to the story, adding VFX expense for no good reason. Avoiding those unnecessary green screenshots can have a huge saving on the VFX budget but the main team in control of that is the camera department and the DOP.
In pre-production I did an estimate of the number of green screenshots based on the script which came to 141 shots. I also asked the DOP to provide an estimate which came to 102 shots. Simply having that cooperation and upfront awareness of the expense of each VFX shot meant that the DOP in the back of his mind could avoid unnecessary green screenshots ultimately saving many thousands of dollars in VFX. The green screenshot count in the final film was incredibly close to the DOPs estimate, within 2 shots.
FXG: The Time Lab set was extensive, and VFX really built on top of the work of the Art Department in this respect, didn’t it?
To get the best possible film from the time and budget available, it was incredibly important to have all the departments involved in pre-production meetings about scenes where VFX may be used. A perfect example was a discussion of how high the walls of the set should be for the construction of the Time Machine Laboratory set.
Higher walls mean more cost and work for the construction team but it gave more scope for the camera department to shoot wider shots without seeing the top edge of the set. There was also a suggestion that the edge of the set could be fixed by VFX in post-production by making it fade off into the darkness. For me, this was not a viable solution due to the sheer number of shots filmed in the Time machine lab. Even doing a small number of visual effects on 100 or 200 shots could’ve used up $100,000 or $200,000. Before long an elegant solution was offered by the art department to create a railing along the top edge of the set giving the illusion that the roof of the soundstage beyond the top edge of the set was actually part of the Time Machine lab. I don’t know how much it cost the art department to put plastic piping rails around the top of the set but I’m pretty sure it was less than $200,000.
Personally, those kinds of meetings with all the different departments, are a huge learning experience for everyone and to me, it’s the most rewarding part of the job. It’s the way that we get the best possible film for the time and money available.
It’s incredibly important to have someone from the VFX department in those pre-production meetings to flag issues because a five-minute conversation can literally save hundreds of thousands of dollars
FXG: What was your focus on set?
Since the 1990s I have been a visual effects artist often doing the 3D work and compositing work on projects. So when I go on set I am always working for the VFX artist who will be using the source materials that I provide for them. I put myself in the shoes of the compositor who needs to key the green screen, I imagine myself 3-D tracking the scene using the tracking markers I provide, and I know what it will be like for the lighter using the HDRI’s that are captured on set.
There is a huge difference between VFX materials that are adequate, compared to VFX source materials that are a pleasure for the artist to work with. I am always thinking from the point of view of a VFX artist who will receive the VFX materials. I consider the delivery of high-quality assets to be the most important part of my job as the onset VFX supervisor. I test out the VFX source materials while creating the post-vis shots so I can be sure that I am providing useful source materials for the final VFX shot.
FXG: How did you shoot the destroyed city
For the Destroyed city, we evaluated a range of locations. One option was to shoot in the forest and insert destroyed concrete walls as practical props and VFX additions. The other option was to shoot in a destroyed city set, and extend the scene with VFX. The final film had a mix of both techniques.
As luck would have it, behind the Adelaide studios where we were filming there was a construction site with piles of concrete rubble that was perfect for our Destroy City shoot location. Unfortunately, workmen were taking our precious chunks of concrete away in dump trucks. So production had to frantically call them and beg them to leave the concrete debris in place until we had a chance to film the scene. The big square building on the right is sound stage 1 at the Adelaide film studios where we were filming other scenes.
For practical reasons, I pitched the idea of filming this scene in the car park behind sound stage 1. That allowed us to make use of the existing cyclone fencing as a support for the bluescreen which saved time and money compared to building scaffolding to support the bluescreen.
This solution also gave me the unique opportunity to capture the VFX backgrounds in the construction site at the same time that the live-action foregrounds were being filmed in the car park. I set up props in the construction site with a row of rocks carefully laid out to act as a join between the foreground to the background.
Whenever we were filming a new foreground take in the bluescreen set, I would radio to my VFX assistant Josh, who was capturing HDRI lighting and photoscans of the VFX backgrounds in the construction site at the same time. That way, we were able to make sure that the lighting in the VFX backgrounds was an exact match to the foreground shoot lighting. This was particularly important on the day because the clouds were passing over causing the lighting to change frequently.
My next step is to create the first version of the 3D scene in post-vis and make adjustments until the edit is locked. My postvis setup then becomes a very tangible brief for the VFX vendor to take the shot to final.
FXG: How about the clifftop reveal scene(s)?
On this waterfall recce, I was dropped in the deep end when the director handed me the remote control for the drone camera. My job was to keep the camera centered on the top of the waterfall as Seth flew the drone. For me, the remote controls felt backward, so I am not really sure how we managed to get this test shot. Initially, there was a discussion about filming the actors at the top of the waterfall, but as everyone was understandably nervous about that, I suggested that we could do it as a digital double shot. It was the perfect shot for digi-double because the actors just had to stand still. When the actors took a break from another shoot, I took the opportunity to take a photoscan capture of them standing in a field with the light coming from the right direction.
This lookout scene is revisited several times in different forms throughout the movie, and I used the test footage as the basis for the 3D scene during the development phase. For the final shot, the waterfall was filmed again at ‘magic hour’, presumably by people who had more experience with a drone camera than me.
FXG: Given there was not a huge budget can you discuss creatively solving a couple of shots – such as the burning man. I assume you did this with comp not 3D?
The protester setting himself on fire was an important worldbuilding shot at the start of the film. The problem was that the stunt man’s head did not catch on fire and his skull cap and face covered in fire retardant goo was really obvious in the center of frame. We only had time for one take, so the director asked me if I could add flames in VFX. My first vision simply pulled flames from other parts of the stunt man’s body and cloned them into his head. I didn’t think that was good enough so I decided to do an element shoot, which turned into five-element shoots getting more and more elaborate each time until I was satisfied.
I analysed the head rotation of the original footage and used my arm as a motion control rig to rotate a manikin head that was covered in burning glue. The head had to hit a specific angle at specific times so I attached laser pointers to the mannequin head, and corresponding targets mounted around my shoot area. So if I managed to point the laser beams at the targets in the correct order, then I knew I was getting all the correct head angles. To get the timing right I made a verbal chant for each section, to help me hit the correct angles at the right times.
The alternative may have been to add computer-generated FX flames, but I feel that using real flames gave a lot more detail and realism. It was also a lot more fun to solve this using real flames.
FXG: Did that apply at all to the explosion shot? I guess that was a stunt double?
The shooting schedule on this day was even more compressed than usual because it had been raining all night and the trucks could not make it down the muddy path to the shoot location. So we spent the morning hauling equipment by hand.
For the head replacement part of this shot, I felt it was important that the actor and stuntman were filmed on the same day in the same location with the same camera and getting pulled back the same distance through the same dappled lighting. Without all those things I feel like the head would not match the body very well.
I also had several assistants yanking on wires attached to the trees when the stuntman was pulled back, so that it felt like the whole scene was getting affected by the ‘explosion’. This head replacement is one of the shots I completed to the final comp.
FXG: What were some of your other favorite ways to ‘cheat’ a shot?
To get this spinning lighting effect one option discussed was to spin a light on a beam, but we ended up deciding to keep the light still and spin the actor and camera together. Our Key Grip, Matt Richardson, managed to get his hands on a heavy-duty turntable which was perfect for this spinning spaceman shot and I also used it to rotate the actors for digi-double 3D scanning.
This set extension shot was filmed in Port Adelaide. I had LIDAR captured onset which was really helpful for setting the scale of the flying vehicle and doing the postvis animation.
FXG: Can you discuss the various Holograms?
Some important moments of the story were told using holograms. In addition to filming a hologram with the Alexa camera, I had help from Michela Ledwidge of Mod Studios to use two Microsoft Kinect cameras to capture 3D point clouds of the actors at the same time. The point cloud generated is very much like a 3-D hologram. I hope all cameras in the future will have a depth pass as one of the color channels as really makes it easy to line up the A and B plates.
One of the challenges of filming and background without a hologram is making sure the eye-line matches. Our Production Designer Jacinta Leong had a great suggestion to use helium balloons for planning the eyeline of the hologram woman in the main foyer.
FXG: What has the response been like to the film?
2067 was the opening gala film for this year’s Adelaide Film Festival. When I watch the film, it is such a rich experience, because I am so familiar with all the decisions that went into it, all the things that were added and all the things that were removed, so it’s impossible for me to watch it with any sense of perspective. All the feedback about the visuals that I have seen online has been positive, so that’s very rewarding to see.