Director Scott Charles Stewart’s Priest envisions an alternative post-apocalyptic world pitting man against vampire, and spiritual warriors helping to win the war. The effects in the film range from the vampires to whole cities and towns, a baron wasteland and a train battle sequence. We take a look at just some of the 750 effects shots.
Stewart looked to his former Orphanage colleague Jonathan Rothbart as visual effects supervisor, Jenny Fulle as visual effects producer, makeup artists at KNB Effects Group, and a host of visual effects facilities to bring the Screen Gems film to life. These included Tippett Studio, Spin VFX, Svengali FX, Spy Post, The Senate, Kerner Optical, Iloura, Gradient Effects, The Base Studio and Zoic Studios.
Priest tells the story of a centuries old war between humans and vampires (told in an animated prologue directed by Genndy Tartakovsky) that is re-ignited when the once-idle creatures find resurgence in a new leader called Black Hat (Karl Urban). Priest (Paul Bettany), whose niece has been kidnapped by the vampires, and Priestess (Maggie Q), part of a warrior group who had previously helped humans win the war, must return to continue the battle, along with Hicks (Cam Gigandet), the boyfriend of Priest’s niece.
Rothbart and Stewart relied on both storyboard artists and then previs created by Pixel Liberation Front to realize certain sequences. “One of the important things for the film was that we wanted to make sure our world was very grounded,” says Rothbart. “Even though it’s an alternate world, we didn’t want it to feel fantastical. So while we were doing the boards and previs, we wanted to make sure we had as much interactivity and do as much as we could while shooting even though we knew we’d be adding, say, CG creatures later on.”
The main filming locations were in the desert near Edwards Air Force Base in California for exteriors and on Sony and other sets in LA. Although it was shot (by Don Burgess) in anamorphic, Rothbart felt the particular qualities of the lenses kept the grounded feeling of the created world. “When we were shooting,” he says, “whenever we were in tight shots and were close, we always wanted to be very wide in our lenses. Again, that meant there was a lot of distortion and chromatic aberration, and also all the fun lens flares that come from anamorphic. But it was also fun to have to re-create all those inconsistencies and anomalies in visual effects.”
A new visual effects production paradigm
Visual effects producer Jenny Fulle co-ordinated the efforts of the visual effects facilities, as well as the 3D conversion of the film, through The Creative-Cartel, her newly-formed production service company. Having cut her teeth for 11 years at Sony Pictures Imageworks as executive vice president of production and executive producer, Fulle founded The Creative-Cartel to specialize in assembling VFX and animation teams for specific projects.
“With 750 shots,” Fulle says, “we had to we basically go through and categorise them – there were primary creatures, secondary creatures, digital environments, 3D hard surface modeling and lots of comp work. We picked a few companies qualified to do that work from each of the categories and added companies as were needed, and freelancers as well.” Fulle notes that co-ordinating so many facilities necessitated strong information and asset management, on a world-wide scale. “We spent a lot of days in cineSync sessions with the vendors, hitting our UK vendors in the morning, Australia at the end of the day and people in North America in between.”
The visual effects effort also segued into the 3D/stereo conversion, which was handled by a number of vendors. “We were able to finish our visual effects before we started on the conversion process,” says Fulle. “Now days, tracking and managing the conversion process is as difficult as any big visual effects project. Suddenly every single shot in the movie becomes a visual effects shot, so every single cut in the movie turns into a shot that has to be tracked. It’s one of the things most underrated by the conversion facilities and the people taking it on.”
The new paradigm, as Fulle sees it also, is to be able to mould to the filmmaker’s needs, as The Creative-Cartel is currently doing for Ghost Rider: Spirit of Vengeance and Ted. “We really don’t have a particular process,” she says. “It’s about what set-up would work best. Do you want to all work closely together? Are you comfortable enough that you can do a lot of this stuff virtually? We can provide the infrastructure and expertise to make it possible.”
Vampires with no eyes
Part of that new approach to the visual effects production was reflected in the creation of Priest’s vampire drones. Creature designer Chet Zar designed the vampires, with a model then provided to the principal vampire vendors, Tippett Studio and Spin VFX. “Scott and I talked early about the look for the vampires,” says Rothbart. “He didn’t want them to be guys walking around in smoking jackets. He wanted them to be creatures and be scary. They had to have some similarity to humans because they both evolved in this world, but he didn’t want them to be guys in suits.”
“In terms of design,” Rothbart continues, “we started talking about how they live underground, and their eyesight and how they hear things, and that brought us to them not having eyes. It actually gave us a really unique look. Then for the motion, I’m kind of a reference nut, so I drive the VFX guys crazy with constantly throwing different animal reference to them of how they would move. I loved how gorillas and monkeys stood on their hind legs and could walk, but really used all four of their limbs when they’re doing power moves. Then as we were working on their motion, I started looking at large cats and tigers and cheetahs. There was such an elegance to the way they ran. When you start breaking it down, you can see the weight and the power in the way they move, although it was still very much a spidery motion.”
Priest and Hicks encounter the creatures at Nightshade, a vampire reservation, which was shot at the Fort MacArthur gun battery in San Pedro. Here, production relied on shooting with vampire dummies and stuntmen in gray suits for reference. “We started shooting nights with a lot of action and back and forth with stunts,” recalls Rothbart. “Paul Bettany really did a lot of his own stunts. We built that sequence to have as much hand-to-hand combat as possible, which was very visceral and could be ugly and messy in its look and feel. So it was shot very hand-held and became a very physical sequence.”
Later, Priest, Hicks and Priestess discover a hive where the vampire colony had been living. Expecting it to be full of vampires, they instead find only one remaining creature – the hive guardian. A beefed-up and brawnier version of the drone vampires, the hive guardian was designed from the ground up by Tippett. “They wanted wanted to take the drone design and see what would happen if this guy had been working out all day and was super-strong,” says Scott Liedtka, a co-visual effects supervisor with Blair Clark for Tippett. “So we ended up adding a harder shell to his back, with very leathery skin, almost like a rhino, and then bulked up his chest and arms to give him a really imposing strength. When we posed him out and had him running around, he opened up his mouth and had big teeth and lots of slobber.”
Tippett art director Mark Dubeau worked up a number of variations for the hive guardian, which was then modeled in Mudbox, although based on the drones. “We had a muscle system for both the drone and the hive guardian,” says Liedtka. “We did scapulas, but not a lot of other bones. The guardian was so muscular that it was mainly jiggle in terms of movement, and his back was this tough hide so there wasn’t too much bone sliding there.” The hive guardian’s skin, and the skin for the drones, was realized with sub-surface scattering and a layer of goo in the skin shader. “The wet goo layer on the drone is actually a real story point,” says Liedtka. “They use that to know there are drones around, like footprints. And then in the guardian sequence, especially, all of those wet surfaces pick up highlights and gleans and glints.”
Tippett created a full face rig for the guardian, even though the creature did not deliver any dialogue, focusing particular attention on the tongue. Shots were composited with a mix of Shake and Nuke for environment projections, and things like dust, scrapes and dripping slime sometimes came from practical elements shot at the Tippett stage.
Says Rothbart: “The hive guardian sequence was very physical and visceral again. For the real interaction, for one of the stunts where Paul is hit by the hive guardian, we built what the stunties like to call a ‘soap bar’ – which was this huge rectangular grey shape that had bars coming off of it. We would have the stunt guys run it on a rig and it would slam into Paul and he would roll over the back of it. Then we would put him on a wire and spin him. It made it a lot easier to get a real interaction.”
“Tippett had to do a lot of digital set extensions for the hive guardian fight,” adds Rothbart. “We had to shoot a lot of our shots in the one direction, because we only had one side of the cave, so even though they were bouncing around to different areas as they were fighting, we would always end up shooting in the same direction and re-work the rocks and bits and pieces. When Priestess is running along the upper areas of the cave being chased by the hive guardian, we were blowing up air mortars and set pieces as much as possible. We just wanted the hive guardian to feel messy, as if he is solely focused on slaughtering these people in his lair.”
Additional vampire drone shots were added to the mix by Spin VFX, under visual effects supervisor Jeff Campbell. The studio also completed shots of the Queen drone and drone pods. After receiving the drone and Queen assets with textures already created, Spin developed its own rigs and shaders and integrated the creature work into their existing VFX pipeline for vampire nest shots, the train sequence and other scenes.
To create the drone pods, Spin carried out some tech development to generate crowds and the squirming movement for within the pods. “The tech solution for the pod movements was very similar to bone and muscle simulations,” says Campbell, “except that the skin membranes had more elasticity. We used a cloth envelope with rigid bodies that could be deformed with the vampire drone rigs. Then we would just simulate this solution over the animation of the rigs.”
Maya was used for animation, led by Peter Giliberti, with the pod fluid that gushes out completed with RealFlow. Muscle and skin movement for the drones relied on solid skin enveloping and corrective blend shapes, with rigging lead by Glen Chang. The final shots were rendered in RenderMan.
For one scene, in the Sola Mira hive, Spin was required to work with a large number of creatures. “One shot in particular involved a mass of drones all converging to a victim tied to a pedestal in the center of the hive,” explains Campbell. “To get a sense of the actual work involved I had animation block out different numbers of creatures in a variety of snapshots to present to Scott. The one he picked had 42 creatures in it. We had our work cut out for us. We couldn’t use Massive because we wanted full creative control over all the individual actions. All the creatures were keyframed. Creative changes had a big domino effect since they were all climbing and fighting each other. We would have regular review dailies of each departments published work internally and externally with the client using cineSync.”
Svengali delivers a city of oppression
Svengali FX handled around 80 shots for the theocratic city known as City Five, and the Jericho wasteland outpost. The city, in particular, was designed to feel enclosed, isolated and oppressive. “When [Svengali CEO] Jamie Venable first talked to me about the job,” says Svengali visual effects supervisor Robert Nederhorst, “the reference was to be Blade Runner, but an even grimier version of that world. It’s actually daylight in there, but the reason it’s so dark is that the city’s pollution has created a dark cloud. A lot of the design sense was to make things dirty and grimy – make them look they have been kind of built, then they ran out of money, and so then they needed to put pipes on the outside – details which would make it look much more organic – like wiring connecting buildings together.”
Production provided Svengali with some limited concept art that matte artists took further to create soaring 600 meter buildings and pockets of detail. “It went from broad concepts of cities to actually going building by building,” explains Venable. “Jonathan Rothbart would circle different areas in our concepts to say that he liked that particular zone, and from there we built a toolbox of different buildings.”
To make up the city, a Svengali layout artist would set out the scene almost like a previs shot. The studio would then build six to eight blocks of fully modeled CG city parts. “Then we’d take the front two buildings that you would see,” says Nederhorst, “and trick those out like crazy once previs was approved. That would give scale to the buildings. Then matte artists would come in and add additional buildings in the distance with matte paintings and projections. The enormous cathedral, for instance, is a projected matte painting.”
Svengali modeled geometry in Maya and sometimes in 3ds Max, using Photoshop for paintings, with lighting and rendering done in V-Ray and Cinema 4D and Nuke used for projections and then compositing. The ‘toolkit’ of buildings and parts could be placed and matched with a proprietary tool.
For the Jericho environments, some of which are featured in harsh daylight, production shot a practical set that Svengali extended. In addition, a large degree of atmosphere and smoke was added to the Jericho and City Five scenes, including shots showing an elevator. “That was a 100 per cent CG building with a projected matte painting at the bottom,” says Nederhorst. “The beam of light was a volummetric effect in 3ds Max. I composited this shot myself, adding some lens flares in Nuke, and was pretty excited to work on it seeing as it was an ode to the Bradbury shot from Blade Runner.”
Otherworldly matte paintings by The Senate
The Senate contributed just under a dozen shots for traveling scenes of Priest and his entourage moving through the desert wasteland. “We did things like expand the desert, put a town on the horizon, introduced a sand storm,” says The Senate visual effects supervisor Richard Higham. “Apart from those shots, we also created several matte paintings, one a full 3D projection, to help sell the story of the key locations and show a passage of time as they travel.
For one series of shots, Paul Bettany was filmed on a motorbike in front of a greenscreen, with The Senate placing that element in front of three different matte paintings of rocky desert outcrops, ranging from a morning setting to late evening. “Editorially, they wanted to keep the bike moving in one go but with the background fading from one to another to suggest how long and far he had been going,” says Higham.
Artists created the paintings in Photoshop, also adding dust for the bikes using a particle emitter in Maya. The bikes themselves were also extended in CG. Another shot featured a pan up from the bikes to reveal a large mountain in the distance, created as camera projection based on rough geometry in Maya. “I think for us,” says Higham, “the interesting things about our shots was that because it’s quite a stylistic film, they needed to create a backdrop that was a little bit otherworldly but still feel like it was actually photographed.”
Spy Post gets on board the vampire train
Discovering that the vampires are using trains to travel, Priest, Priestess, and Hicks set out on the trail of Black Hat. For scenes of their approach, Spy Post completed 125 shots featuring a CG train, tracks, two motorcycles environment extensions, a set of massive statues at the entrance of the wasteland, along with dust, smoke, fire, explosions, blood and mist.
Based on previs by PLF and storyboards, production shot a full-scale train in the desert near LA, made up of a semi-trailer pulling three box cars. “The box cars were all on truck wheels,” explains Spy Post visual effects supervisor Michael Janov, “which later would have to be painted out, and replaced with train wheels sitting on a track. But there was a definite advantage of having the truck driving in the desert, as it kicked up dust adding an atmosphere that is always tricky to add later in comp.”
“It was important there was something real out on location for our actors and stunt guys to interact with and that we could use for lighting,” adds Jonathan Rothbart. “The concern is in your fully CG world that if they start playing as CG everywhere, I think the audience loses a lot of the jeopardy and what’s going on with the scenes and then you lose them as a viewer.”
Rothbart laid out parking cones along the path of where the semi would be driving, objects that Spy would later use as tracking data in the sparse desert. “It made for a little more paint work,” says Janov, “but assisted greatly in getting the camera move just right. Some of the camera work was shot from a fast moving vehicle, so any extra info we could get to matchmove would help.”
Kerner Optical also built a 1/12th scale version of a locomotive, tender, fuel car and box car, which would later be blown up. “There was so much detail put into this miniature,” recalls Janov. “They had grease painted in strategic places like around valves, and vents. There was dust collection around the wheels and along the base of the train. We created an on-set turntable where we could prop individual portions of the train, and with as flat lighting as possible – no harsh shadows, no bright reflections – we took pictures of the side, front and top. We then used these photos as templates for our model, as well as a base for our textures.
Spy modeled the train in Maya, which was rendered in mental ray, writing a shader that added in broken up reflections to simulate where dirt would collect, and other areas where paint would have been chipped away and exposed underlying metal. A similar modeling approach was followed to create the CG bikes and bike extensions.
Both dust and the train smoke stack were added to the shots, using a combination of Particle Flow, FumeFX and Krakatoa in 3ds Max. “We created fluid simulations that were easily scaleable, repositioned, or speed-warped based on the shot,” notes Janov. “In comp, using Nuke, we used practical smoke to help break up the edges, and to give it another level of detail that is sometimes difficult to get in CG fluids. In other cases we would favor Afterburn for dust, as it allowed for more flexibility and the render times were reasonable. For some shots, the camera drives through the volume of dust, and Afternurn allowed from some speedy render times, even when we added God rays.”
The desert plates also required some visual effects work to remove background mountains in order to make the environment feel more sparse. “We painted the mountains out, created a ground plane that had a shader matching the look of the desert ground, and with a little bit a massaging in comp, we were able to extend the horizon into infinity,” says Janov. “We also created a matte painting of mountains that looked further away than the mountains in the plate.”
The large statues at the entrance to the wasteland were rendered first as concept drawings, then as basic shapes, before detailed modeling began in ZBrush to show age and weathering. “We would take huge chunks out that would reveal an armature that we built inside of the statue,” says Janov. “Once lookdev’d and in a shot, we did additional paint work on top of the statue to help give it more detail.”
Spy Post also delivered the stereo conversion on its shots for Priest, once all their visual effects work had been completed. “At first we revelled in the convenience of re-rendering everything and putting it back together the same way,” recalls Janov, “but on a couple of shots that actually became a big hindrance. When creating a 2D shot, if elements of an intricate explosion are rendered separately, they can be composited in an arbitrary order. For instance, the smoke can be composited in front of the fire.”
“However,” adds Janov, “when rendering in 3D, where both layers have intersecting depth, that smoke is composited in front of the fire but is at a depth further away than the fire it’s in front of. It creates stereo artifacting and can be painful to look at. In these cases, we had to balance two less-than-ideal solutions: run everything in a single pass and lose the control that compositing provides, or do frame-by-frame paint fixes on the final stereo images. We eventually leaned more toward the paint fixes solution because it kept us more faithful to the approved 2D shot.”