In The Wolverine, Hugh Jackman returns as Logan – this time as an almost broken man and mutant, facing new adversaries and his own fate. Visual effects supervisor Phil Brennan and VFX producer Jamie Stevenson oversaw several effects shops in order to bring the film to life with scenes such as the Nagasaki nuclear explosion, a dramatic train-top fight, a CG Silver Samurai and, of course, Wolverine’s claws. We break down just some of the major shots, sequences and previs work in the film from Weta Digital, Rising Sun Pictures, Iloura, Make-Up Effects Group and Halon Entertainment.
***Note: this article contains spoilers***
Nagasaki’s nuclear fate
What happens: Logan is held captive inside a well at Nagasaki just as a nuclear bomb is dropped on the city.
The challenge: Rising Sun Pictures took plate photography of a Japanese army base filmed in Sydney and added the nuclear explosion and its impact on the base.
“Jim was after a nuclear explosion that wasn’t your typical library footage where you see the big column of mushroom cloud out on top,” says Rising Sun visual effects supervisor Tim Crosbie. “You actually see it starting from scratch. We built this pyroclastic cloud – they can travel up to 400 miles an hour, roughly half the speed of sound. We looked at the natural phenomena of what comes out volcanoes and compared that to how the destruction clouds of nuclear explosions tend to work. It was quite similar.”
Production filmed in an area of southern Sydney made up to look like a Japanese army camp. Rising Sun replaced the city on the horizon with Nagasaki views. “It was primarily matte painting but we also ran some geo underneath it so that when the explosion went off all the fluid sims and particle sims and destruction effects,” notes Crosbie. “It was a combination of deconstructing the location that was there, so that we could then re-build it back up again. So all the huts, guard towers. The whole location was LIDAR’d so we could reconstruct the area. And then we destroyed them, we blew them up using a combination of smaller explosions to make a larger one.” RSP relied on the Bullet solver inside Houdini 12 to create the explosion effects.
One shot features Logan doing a last minute jump into the stone well, having saved a Japanese soldier named Yashida. “It was the surfing the wave moment,” recounts Crosbie. “Initially there was just the camp as shot, with Sydney in the background off-screen on the horizon. Hugh was shot against a bluescreen on wires being lifted up and dropped into the well. It basically meant we had to almost re-build the entire frame. The only stuff we really kept was the well head and the foreground and mid-ground ground plane, but even that was pretty much covered with the particles running towards camera.”
Logan and Yashida shelter inside the well, ducking debris and fireballs. These were filmed as practical fire elements by Phil Brennan and then composited into the shots. “When the flames come past camera that was a combination of practical and CG fire,” says Crosbie. “It helps break up the smooth lines that can sometimes come from fire sims.”
What happens: After Logan bears the brunt of the Nagasaki nuclear explosion by shielding Yashida from the blast, his severe upper body and facial burns are shown healing.
The challenge: Iloura crafted the healing shots using complex tracking, CG and compositing solutions, as well as growing back Wolverine’s signature hair.
“For ‘Crispy Logan’ as we called it, one of the biggest challenges was to make it look like his skin was healing, and not just dissolving back to his healed skin,” says Iloura visual effects supervisor Julian Dimsey, who also oversaw work with VFX supe Glenn Melenhorst on the show.
Practical make-up effects for Logan’s burns were created by Make-Up Effects Group under the supervision of Nick Nicolaou and Paul Katte. The studio sculpted burn shapes and wounds that were converted into silicone and vinyl prosthetics to simulate the woulds. For the headling shots, Jackman performed with no shirt on and some lightly drawn tracking markers. “We used a layered approach for the tracking,” says Iloura lead lighter Drew Wood-Davies.
“We had a highly detailed digi-double scan of Logan given to all the vendors. We rigged this and turned it into an animatable asset. That was roughly matchmoved to the shot to the point where we were happy with it. Then our lead creature guys went through and picked a series of key poses we could create high-res blend shapes for. We then used those to correct the matchmoves.”
The next step was some painstaking soft tracking done in SynthEyes on Logan’s skin to match muscle movements and twitches. “All those points were then projected back onto his surface in screenspace,” says Wood-Davies, “and then skinned to give us all those little subtleties in the movement. Then we could move into lighting and shading and put all the effects on top of what was a pretty solid track.”
The CG components of the shot were created in Maya and rendered in V-Ray, with compositing in Nuke. “The burn effects were animated mattes generated in Nuke using a series of library elements and animated masks that would control the reveals of different parts of the surface at different times,” explains Wood-Davies. “The hair was handled in a similar kind of way – we used Yeti within Maya, and used animated mattes to handle the re-growth of hair in those areas once the healing effect had taken place.”
“We also used those masks to control displacement and things like that so as the effect retreated it didn’t feel like it was dissolving,” adds Wood-Davies. “The masks got us so far and there was also some comp’ing down the track to make it work.
Logan and the bear
What happens: Logan is seen walking down the mountain on which he is ‘roughing it’. A large bear walks not far away. Later, the bear is seen near death and Logan makes the hard decision to put it out of its misery.
The challenge: The walking bear was a Weta Digital creation that was based initially off the animatronic bear built for the later nighttime scene by Make-Up Effects Group and a real grizzly bear in California that was photographed for reference.
The digital bear was modeled in Maya and its fur groomed with wmWig, Weta Digital’s latest incarnation of Barbershop, its proprietary hair modeling and sim software. “Muscle and skin was rigged using wmTissue, our in house solver,” says Weta Digital visual effects supervisor Martin Hill, “then textured with Mari, shaders and fur in PRman, using our custom shaders based off Zinke’s dual scattering methods with a spherical harmonic approach for image based light for scattering and diffusion. The close-ups were done in the same way for the non-two and a half D parts, with extra wetness on the fur shading and groom.”
Make-Up Effects Group, under the supervision of Nick Nicolaou and Paul Katte, built a 12 foot tall animatronic bear, that was used to shoot the dying creature after it has been hit by a poisoned arrow fired by hunters. Logan comforts the bear and eventually ends its suffering. The bear was controlled by three puppeteers on location articulating the muzzle and paws via remote – in simulated pouring rain and mud.
Construction began with a maquette sculpt for approvals. The bear head, paws and teeth with sculpted in clay for fine detail – the teeth fabrication went even further with moulds taken to craft dental acrylic versions. Eyes were formed by moulding a ball-bearing, creating a resin copy and grinding down the top layer of the resin ball to represent an eye and painting in a clear resin lens look. The bear’s body was made up of a muscle suit with a woven hair fabric covering and individually hand-punched and implanted hairs.
The pin bed
What happens: Logan visits Japan to see a now close-to-death Yashida, who is resting on a ‘pin bed’ that contours to his position and movements.
The challenge: Iloura produced digital pins to match a practical version, with the actor shot mostly lying on a greenscreen block element.
“We started with a matchmove of the actor’s movement and then we had a rig of the pin bed with all of the pins skinned to a surface that would conform to his motion,” says Drew Wood-Davies. “That would give us a rough guide of how those pins should conform to support his movement. Then on top of that there was a secondary sculpt, and there was a bit of keyframing and manual sculpting too. They also wanted micro movements as well, hard to pick up in matchmove, where if he twitched his hand or his neck shifted ever so slightly, they wanted to convey that motion in the pins, then all that was keyframed in Maya as another layer of motion. It was rendered in V-Ray.”
“A lot of the interactivity with Yashida was done as a 2D solve in the end out of Nuke,” adds Julian Dimsey. “The lighting department gave out a bunch of passes to allow the compers to displace and warp Yashida’s skin and his pyjamas based on the movement of the pins. All of that interactivity was generated as a 2D effect.”
Production had made a real prop with static pins, so HDRIs from on set served as reference photography for the metallic pins. “We made a model of the bed,” notes Wood-Davies. “Then we would project Yashida back onto the digital geometry so we would have him reflecting back on the pins and occluding in the correct spots.”
Tokyo drift at 300 km/h
What happens: Wolverine follows Mariko (the granddaughter of the now-deceased Yashida) onto a Bullet train but soon comes face to face with her attackers, the Yakuza. They fight and eventually end up on the roof of the speeding train as it zooms through the city, dodging signs and gantries along the way.
The challenge: Staging such a dramatic train-top fight sequence was clearly going to be too dangerous for the principal actors and even stunt doubles. Instead, backgrounds were captured with a special multi-camera rig from a Tokyo freeway, then engineered by Weta Digital into live action greenscreen plates of the fights, along with digi doubles and digital environments.
“The first thing production did was phone up Japan Rail and say, ‘Hey, look, we want to do this sequence.’, recalls Martin Hill. “The first thing that came back was that, ‘The tops of our trains have 25,000 volt pantographs on top – do you really want your actors on top of that?’ Well, basically, no.”Watch part of the final sequence.
It was decided, then, that actors and stunt performers on wires would be filmed on a set piece surrounded by greenscreen. For what would be viewable beyond the train, Hill notes that the filmmakers wanted to see a lot of depth in the backgrounds, not just a few CG buildings. But the sequence would span nearly two and a half minutes of footage – a significant city build for that length.
That led to a decision to shoot backgrounds practically as much as possible, both for actually filling out the sequence and for reference. “The overall VFX supe Phil Brennan had this great idea,” says Hill, “where he said, ‘Why don’t we look at filming, rather than on a train, but on an elevated freeway that’s going through Tokyo.’ So we scouted some locations and found this elevated freeway in Shinjuku. We can drive up and down that all day and get lots of plates.”
However, at that point in the production the actual shots had not been designed, so rather than plates the team adopted a ‘Google Streetview method’. “But instead of having a big panoramic cam on top of a van, we built a rig that had eight 45 degree angle RED EPICs that gave us massive resolution driving down all the massive lanes of the freeway,” explains Hill. “We let a bit of air out of the tires of the van and kept a constant 60 kilometers an hour. So if we shot at 48 fps we just needed to speed up the footage by 10 times to give us the 300 kilometers an hour required.”
The results were moving backgrounds filmed from an elevated position to approximately match the position of the train if it had been filmed for real. “We still had to take that footage, line it up for all the shots we wanted to compose to go with our greenscreen backgrounds,” says Hill, “and then do a lot of one point perspective and two point perspective photogrammetry techniques to re-warp those plates and give them the depth and the nodal point offsets we couldn’t get by merely taking the footage and placing it as a background. ”See how the greenscreen portion of the train fight was filmed.
A major effort was then launched for adding extra components into the shots, things such as roads with moving cars, rush hour traffic, train tracks, gantries, signs, billboards and, of course, the train itself. “Jim Mangold was very keen to make the city very Japanese,” notes Hill. “We showed him the original footage, sped it up and showed it to him and he said it looks great and fast but if you say ignored the bit of text from the car registration plates, you could be almost anywhere. So we had to add lots of colorful billboards showing all manner of noodle bars and train signs.”
“And to give it a sense of speed you always want things much closer to camera, so they’re really whipping past,” adds Hill. “So we were lowering all the heights of the gantries so that they would whip over the actors and the camera and give it a sense of speed as the trains go through it. It became a story point as they’re always trying to avoid these gantries – ducking underneath them, interacting and leaping over them.”
Weta Digital retained close-up buildings captured with the multi-cam rig, although some were then re-projected onto geometry in Nuke to “give us that two point perspective augmentations that we needed to be able to move the camera from where it sat on top of the van,” says Hill.
The back plates were solved first before Jackman and the other actors were filmed on set. The established environment was a little overcast without full sun which the DOP could then match on the greenscreen stage. “We’d have a keylight replicating the sun,” says Hill. “There was a giant propeller on a cherry picker with the 10K lights and one of the grips would just revolve it at a certain speed so you’d have this light whipping over the background. It was however difficult to get a practical propeller to line up and cast a shadow exactly when Hugh ducks to avoid a gantry. In shots like that we didn’t run any kind of shadowing and just did it in post.”
Ultimately the practical fights and wire work were augmented with digital doubles, especially for some of the more dramatic scenes of the actors leaping to avoid gantries.
Scenes were rendered in RenderMan. “We ran a very large environment but kind of a rolling one,” explains Hill. “So for each shot we’d take about a kilometer of track with more or less a standard setup. The first thing we do is render it, see how it looked, judge the speed, or add more drama, adjust the height, the spacing between the gantries. Then we did shot-specific things like if there was a particular gap in the background that had a lovely view to the city we’d try and not block that with a giant billboard, or if there was a particularly nasty looking building then we certainly would.”
“We also had to think about how are we going to deal with extra bounce lights from buildings that don’t exist – say red buildings – and how to represent those back onto the train or onto Hugh. They do get fed into an IBL but there’s also an element of over-cranking it. We wanted to art direct the lighting and it had to have a certain impact.”
What happens: A robotic medical beetle is implanted inside Logan’s body. When he realizes the parasite is limiting his healing powers, he cuts open his own chest to remove it.
The challenge: Iloura brought the aggressive tentacled beetle to life both in an appearance in a jar of liquid and as Logan pulls a beetle out of his chest.
“The beetle was modeled in 3ds Max, with a ZBrush sculpt and rigged in Maya,” explains Drew Wood-Davies. “The tentacles had a keyframe animation component as well as a simulation component. They wanted it to look very aggressive so we had to convey that in the animation.”
On set, Jackman held a small metal prop to simulate its scale as it holds the retrieved beetle. “We did a clean-up pass on that and then add in our digital creature,” says Wood-Davies. “We had to layer things in like blood to make it all sit with the stuff going on.”
For the beetles seen earlier in the jar, Iloura ensured that the caustics gave a small sense of interaction. “One of the things we did was have accurate caustic occlusion for the shadows of the beetles on the table in the foreground,” notes Wood-Davies, “which I felt were quite enjoyable and helped sell the shots.”
Wolverine versus Shingen
What happens: After removing the parasitic beetle, Logan fights Shingen who has also been battling Yukio. Shingen is eventually slain.
The challenge: Rising Sun carried out sword and Wolverine claw replacements for the fight, as well as background composites and rain.
“Shingen had just the sword hilt and we replaced it with a CG sword, including for a shot of him plunging it into Logan,” explains Tim Crosbie. “Sometimes you can find if you don’t have the weight of a full sword and you’re miming pushing it in and out, well, the force that actors try and tend to put on with shaking arms make it feel like there’s a fair bit of force there. So we had to rebuild the plate behind and stabilize the hand and the sword to sit properly on the sword as it’s withdrawn from Wolverine.”Watch part of the fight sequence.
Jackman filmed his fight scenes with different sets of claws, depending on the amount of action and interaction required. These were: plastic claw, rubber claws, metal claws and, finally, small two inch long stubs.
“Those metal claws are pretty sharp,” observes Crosbie. “You can put an eye out pretty easily – they are polished aluminum. They had to be sharp otherwise they wouldn’t play properly in close-up. But only a certain level of stunts can be done with them, beyond which even the slightest bump can cut his knuckles – so that’s where CG takes over.”
Multiple still shoots of the claws against black were carried out for reference, and HDRIs also taken. “We pulled all that into our asset system,” says Crosbie, “modeling them in Maya and after it was modeled we did the textures, then based the lighting on HDRIs. We’d take a still frame from a plate as the background and spin the claws in the scene, sit them next to the live action ones and see who could tell the difference.”
What happens: En route to Yashida’s lab, Logan is ambushed by Harada’s ninjas and battles them in the snow, eventually being pinned down with multiple cabled-arrows and the effect of Viper’s venom.
The challenge: Filmed in summer time at Sydney’s Olympic Park, Rising Sun extended the village set and added snow, digital arrows and cables.
“It really was a wonderful set built in Sydney Olympic Park,” says Tim Crosbie. “We were doing night shoots but it was shot in Sydney in the middle of summer. It was quite a lot of fun, but surreal running around in t-shirts and shorts when it’s meant to be the middle of winter in snow.”
The set was LIDAR’d, with RSP carrying out extensions via camera projections onto geo and adding mountains in the far background plus extra snow. Wider views back towards the village from the lab were also realized as matte paintings. The studio handled wounds to Logan’s face and claw effects, as well as the green venom into which Harada dips an arrow.
For the shot of Logan being bombarded with multiple arrows on cables, Jackman was filmed with just a few props attached to his back. “We had practical arrows in his back with a few ropes hanging off him,” states Crosbie, “just to get a good reference for how it should look. Jim wanted certain looks referencing back to the comic framing. The ropes were done using nHair in Maya to get the right flexing and movement.”
Shedding some skin
What happens: Amidst the lab battle taken place between the Silver Samurai and Wolverine, Viper engages in hand-to-hand combat with Yukio. At one point she sheds her snake-like skin.
The challenge: Weta Digital produced the effect of Viper ripping off her peeling skin.
Actress Svetlana Khodchenkova mimed the peeling off action. “Unfortunately there was no object for resistance for her peeling skin,” says Martin Hill, “however Svetlana did a great job putting some weight into the performance. Before we started any effect, we re-timed and warped the plate to add more tension to the performance, then worked from that.”
Khodchenkova was also filmed with several markers on her skin but, as Hill notes, “getting a perfect match under the neck, on the platysma especially was difficult. Once we had that working, we used a combination of two and half D techniques to apply textural and lighting layers onto Viper’s skin.”
A shiny Silver Samurai
What happens: Wolverine confronts a Silver Samurai at the lab made out of adamantium and designed to draw out the mutant’s powers of healing.
The challenge: Weta Digital brought to life a nine foot tall fully reflective fighting ‘machine’ that engages in hand-to-hand combat with Hugh Jackman.
A 3D model of the Silver Samurai had been 3D printed by production and chrome painted using electrolysis methods for use on set. “He looked fantastic,” noted Martin Hill, “but unfortunately he didn’t move, so that’s where we came in.” In addition to creating a ‘living, breathing’ version of the Samurai, Weta Digital re-worked the original design to ensure it would perform the necessary martial arts moves. “The design that had been worked on was very graphic and had a strong presence but it didn’t necessarily have all the movements that Jim wanted him to perform,” says Hill. “For example, he had a very wide chest which would make it difficult to perform traditional Japanese moves, say holding a katanas with two-handed moves.”
Stunt performer Shane Rangi stood on stilts during filming as the Silver Samurai. “We did dress Shane up in a pseudo mocap suit with witness cameras so we would be able to capture his performance,” says Hill. “But post-shooting the motion moved much more towards martial arts style and more traditional Japanese style and so we were able to do a lot of keyframing.”
Animation wise, the Silver Samurai featured a completely solid face mask, so Weta Digital sought to give the character emotion from body performance. “There’s a point where he stabs his son-in-law and we had to hesitate beforehand – it had to look like he was doubting himself when he did it, with no eyebrows, no face, no eyes,” states Hill.
Perhaps the main challenge for Weta Digital was realizing the Silver Samurai’s highly reflective surface. “He’s pretty much chrome,” notes Hill. “We were worried that he was going to look incredibly digital and that it was going to be very hard to make him look solid and real and not just like a mirrored surface. So we spent a long time replicating the multi-layered structure of brushed metal with a micro-facet solution to give us an anisotropic patina to him, but also give us a mirrored reflection underneath of the flatter parts in between the grooves. We took the original practical reference on set and had our shader writers actually started machining and milling different bits of metal and having a look at different responses from complicated materials.”
The Silver Samurai also carries a large sword that he is able to heat up and cut through almost any surface, including Wolverine’s own adamantium claws. “In the original previs the sword had a purple glow around it and left a trail behind it – almost Japanese anime style,” says Hill. “It looked cool but once we got into the lookdev of it, Jim wasn’t really sure about the ethereal look. We went back to focusing on how katanas are actually made, how they fold them the metal and stick them in to the forge. You get a really interesting heat treatment across the sword while they’re forging them – all the cracked patterns as the layers of the metal are coming away – you can see it’s all based off the temperature of the metal. That’s we ended up replicating. Once they’ve cooled down they have this thin film interference layer which looks a lot like the color patinas you get around arc welding. Combining these things with the damage during the fight – it gave him a much more convincing look.”
To replicate the environment, HDRIs were taken at different heights and served as reference for the final environment maps on the Silver Samurai. “What that generally gives you is a real representation of what the samurai may have looked like in the scene,” says Hill. “But the DOP isn’t necessarily lighting that dead space where the Samurai is, so we really had to fill in the blanks.”
“We designed the lighting more procedurally,” continues Hill. “Because he’s so reflective you can play with parallax in the lighting, so the lights can go behind pillars, say and it’s not just the shadow – you’re actually seeing the reflections of the lights disappearing and appearing – you get real depth.”
Hill says Weta Digital had to light the Silver Samurai in a very graphical way. “You can’t just take a big area bounce light and have it as a crooked angle against those lines – it needs to be complimentary against his structure and the way he’s designed. The lighting needs to be sympathetic to that.”
Eventually it is revealed that inside the Silver Samurai is Yashida, kept alive until he believes he can absorb Wolverine’s powers. As he attempts to do so, the old Yashida transitions back to his younger self – a morph completed by Weta Digital which simulated various stages of the transformation. Full FACS sessions with a three-camera setup were undertaken to acquire facial features and gestures. “We also recorded the full length of the scene with each actor in each make up stage 2 or 3 times with multiple cameras, which gave us precise reference shapes to aim for rather than always rebuilding from FACS primitive shapes,” says Hill.
Artists paid attention to changes required in hair and eyebrows, for example, and the need to adapt sub-surface scattering. “Some areas were more translucent and more leathery,” explains Hill, “so each of those stages were lookdev’d independently and then we’d be morphing in those individual maps in line with the blendshape transition to always keep the character looking photoreal.”
Keeping each point in the transition looking like the actors when blending between was one of Weta Digital’s main challenges. “Some early tests tended to look like a entirely different character which was neither of the actors,” says Hill. “The other challenge was to keep the textural and sculptural blends working in tandem without the effect appearing as a wipe.”
Find out more about how Weta Digital completed the Yashida transitions in this interview with compositor and fxphd prof. Tahl Niran.
Previs’ing The Wolverine
The Wolverine’s previs was handled by Halon Entertainment and supervised by Clint Reagan, who approached the work slightly differently to a typical major comic book film. “Jim wanted it to be a more personal,” says Reagan. “To keep things tighter he didn’t want the super big over dramatic exaggerated comic booky stuff. For example, Jim didn’t want us to animate something so fantastic that say the stunt guys couldn’t even do it, and then it would become an all-CG thing. So we worked really closely with them to do almost like a postvis over their stuntvis work. The stunt guys were able to get just a visceral intensity that would take a whole lot of time in the previs setting. But we could give it so much more context than just in the practical setting of their warehouse.”
Halon took the plates of video footage filmed by David Leitch’s stunt team and edited and ‘postvis’d’ over the top, which would then serve the second unit and stunt teams during principal photography. Mocap was also relied on early on to flesh out various fight stunt scenes. “Fox has a MVN suit from Xsens,” notes Reagan, “so we went down to David Leitch’s stages and they did mocap of some of our scenes that had been blocked out with boards. Then we did some mocap for some martial arts movements. We incorporated a lot of that early on in the Silver Samurai fight, although much of the mocap became obsolete because of the changes required, but it was still great reference.”
Some of the major sequences created in previs by Halon were the Nagasaki nuclear bomb, the train fight, Viper’s skin reveal, the Silver Samurai encounter and various fight scenes including the funeral, village and a bear sequence that ultimately was cut from the final film. Halon primarily worked in Maya to create previs scenes, using Motion Builder for mocap translation, tracking in SynthEyes, editing in Premiere and added various effects in After Effects.
Reagan, who once lived in Japan, was able to draw on his personal experiences to previs certain shots, including those set in Nagasaki. “The previs was influenced by a woman I met in Japan who lived through the blast in Hiroshima and she took me around and told us some horrendous stories that bring tears to my eyes just thinking about it,” recalls Reagan. “I told Jim stories she had told me about how walls had collapsed on her and how there was a bright flash and nothing happened, then there was a gust of wind, and it was really hot. It just made it so real to me because this woman I had known had relayed it to me.”
X-Men Character Likenesses TM & © 2012 Marvel Characters, Inc. All rights reserved. TM and © 2012 Twentieth Century Fox Film Corporation. All rights reserved. Not for sale or duplication.