Trevor Hogg chats with visual effects supervisors Phil Brennan, Martin Hill, Tim Crosbie, Glenn Melenhorst, James Rogers, Bryan Godwin, Viktor Muller; effects technical director Prema Paetsch and previs director Clint Reagan about such things as claws, nuclear explosions, shedding skin, and a Silver Samurai which all appear in The Wolverine. Beware there are spoilers...
“I ran into some people at Fox who thought I would be a good fit with Jim [Mangold] so I arranged a chat with him,” recalls Phil Brennan (Snow White and the Huntsman) as to how he became the visual effects supervisor for The Wolverine (2013). “We got along well and decided straightaway to go ahead and do it.” With a resume which includes the Johnny Cash biopic Walk the Line (2005), Western remake 3:10 to Yuma (2007), action-comedy Knight and Day (2010), and police drama Cop Land (1997), James Mangold makes his comic book movie directorial debut. “Jim is a big comic book fan so he has an understanding of the whole genre even though he has not done a film like that before. Also if you look at most of his filmography Jim is about character and performance.” In order to articulate his ideas Mangold would literally act out what was needed in a scene to his cast and crew. “Hugh [Jackman] was completely on board with this; he was heavily involved all the way through.” The immortal clawed mutant who is lured and betrayed in modern Japan has become the signature role for the Australian actor who served as a producer on the project. “It gave Hugh some freedom from the previous films as he could explore other aspects of the character.”
“We would have a whole range of different kinds of claws that are practical that Hugh could wear,” states Phil Brennan. “Hugh knows by doing it so many times and by being such a good stunt guy what is safe to use for each shot. Whenever we would come to a particular sequence we would have a chat with Hugh about what he was comfortable with wearing. It could be anywhere from the full metal claws to nothing at all. Either it could be all in-camera or all CG or somewhere in-between depending on what was safe. The other thing with the claws is that they’re retracting or emerging from his knuckles those had to be CG. There are some subtleties in their shape which is slightly different from the way they looked in the past.” The emergence and retraction of the claws was similar to past movies. “There are a couple of moments where they act more slowly but in general it happens quickly. You still have to carefully track them onto his hands and the lighting of them needed to match the scene. They’re not a trivial thing to do and there are a lot of shots where the claws have to be CG or partially CG. One hand has to be CG and the other has the real ones. It’s a mixture depending on the stunt.”
Deciding on the amount of green screen required for each scene was not a contentious issue. “There’s obviously a cost trade off,” observes Phil Brennan. “Given unlimited money Art Departments could build almost anything but they don’t have unlimited money. Given unlimited money in CG we could do almost anything but we don’t have unlimited money either. It’s always a trade off between how many times are we going to see this? If it’s going to be three or four shots of then maybe we should build a small set and do a lot of it in CG. If it was going to be a hundred shots then it is worth building a much bigger set and doing less CG because every CG shot costs money.” Brennan explains, “There is a little bit of the bullet train and the rest of the environment was green screen. We shot a lot of practical plates to help build the environment that replaces the green screen. The Yashida estate was a beautiful set that François Audouy [Abraham Lincoln: Vampire Hunter] created and the Art Department built. The entire compound itself was real and all we had to do was replace the background with Tokyo in the distance in the night sky. There was green screen there but the work was fairly simple from a visual effects perspective. The ice village was a large set that we had to make look bigger. We had to cheat at a few places to make seem as though we weren’t using the same piece over and over again. We added more buildings in the distance. There are a number of shots where the foreground is the set and Hugh is running into a street which has been created in CG.”
“We shot this film largely in Australia and there was a certain amount of work we had to do there” explains Phil Brennan. We did try to spread the work around to a number of the facilities in Australia.” Visual effects help was found in neighbouring New Zealand. “Weta handled the bear, the Bullet Train Sequence, and the Silver Samurai at the end. Rising Sun Pictures are particularly good with dynamics, simulations, fluids and smoke. They did the Nagasaki nuclear explosion in the beginning and the whole Ice Fight Sequence with all of the arrows and ropes plus a whole bunch of other fight sequences. We used Iloura in Melbourne; they did all of the [skin] burning and viper effects, and some of the old man in the bed. UPP we needed two or three big shots done and at the last minute. It was environment stuff which they’re good at. One was an establishing shot of the Yoshida house on the cliff. It wasn’t a shot we anticipated so it had to be created from scratch. Shade did all of the things around the Love Hotel Sequence and a bunch of graphics work for monitors; they’re local so it was easy to interact with them. We also used Method here in L.A. for some last minute stuff.” Other contributors included Method Australia and Halon Entertainment which provided the previs and postvis. “We tried not to break up sequences too much. Rising Sun did the majority of the claw work although everybody ended up doing a bit of claws.”
“Fortunately, Phil and I had a lot of time on-set together to discuss the sequences in-depth [about how to shoot and approach the post] which meant we had a great idea of what he and Jim were looking for,” states Weta Digital Visual Effects Supervisor Martin Hill. “We kept in touch via cineSync once or twice a week to keep on top of feedback and directorial changes. Phil's creative and has great ideas for approaching how to get the look required, but is also completely open to ideas that make shots better. It was a really fun collaboration!” Hill remarks, “On Logan himself, we worked on the claws and various digital double versions which were pretty much design locked. For the Logan Transformation Sequence where he has his powers drained by the Samurai, we designed the look and transition of the character. Similarly the tearing out of the claws and regrowth of the bone claws through the torn hands were designed by us. Previously, the Adamantium stubs remained and the bone claws grew through them. Philip and Jim were open to this sort of input. Other than Logan, everything else was fresh. The Viper was an entirely open design that changed through the production. A practical version of the Silver Samurai existed but his motion was designed from scratch, and went through several iterations as Animation Supervisor Mike Cozens refined his personality with Jim and Phil's feedback. We completely redesigned his sword and the 'Adamantium cutting' effect. As with any show there were countless creative decisions being made constantly that had a lot of scope for our input.”
A showcase action scene is the Bullet Train Sequence. “There have been many movies we’ve seen with things happening on roofs of trains and quite often there are fights,” states Phil Brennan. “We did a lot of research looking at what worked and didn’t in other films. A couple of things stood out. We wanted to make sure that we weren’t doing a green screen train move and everything else is CG. That’s an area where it is hard to pull that off and is generally not that successful. We wanted a certain amount of realism in the background. We researched whether it would be possible to get cameras on a bullet train or any other kind of train in Tokyo and that wasn’t going to happen.” An innovative solution was implemented. “Phil Brennan had the fantastic idea of shooting background elements ‘Google Street View’ style, from the top of a van on an elevated freeway in Tokyo,” reveals Martin Hill. “This involved 8 Red Epics mounted at 45 degrees from each other forming a panorama then driving up and down the freeway. We used this as the basis to build the backgrounds. With a good amount of perspective warping and some photogrammetry style reconstruction, we could create backgrounds than lined up to the foreground GS plates. We added digital elements for all end trains, tracks, wires, gantries, and billboards taking care to interweave everything in the same space. This technique gave us a lot of depth and parallax for free with the mid ground and far ground buildings from the original plates.”
Extensive previs was required for the Silver Samurai Sequence. “We knew going in roughly what each shot was going to be,” remarks Phil Brennan. “It was never anywhere exact because you don’t want to take away the spontaneity from the shooting side of things. The design of the character largely came out of François Audouy, our production designer. There were several iterations which had to be approved by Jim and the studio. Early on in that sequence the samurai is not moving so we built a full scale version of it which is used in several of the early shots before he becomes animated.” The works of a legendary Japanese filmmaker proved to be useful. “For the Silver Samurai we had some demonstrations from Kendo practioners and watched a lot of [Akira] Kurasawa!” chuckles Martin Hill. “For the suit we researched the material properties of the brushed chrome, creating practical swatches of different materials for different areas of the suit.” Motion-capture was considered but a traditional method was adopted to give Hugh Jackman something to act against. “The thing is just over nine feet tall. We took a stunt guy whose thing is working with stilts which made him roughly the correct height and he played the Silver Samurai while we were shooting; that meant that eye line wise Hugh was looking at the right place.” Multiple witness cameras captured the action. “We shot a ton of reference [material]. It was great having the samurai built because it served as a lighting reference.”
A complex shot is when Svetlana Khodchenkova (Tinker Tailor Solider Spy) who portrays the poisonous Viper sheds her skin. “We did everything we could on-set with all sorts of witness cameras but at the end of the day it came down to doing some intense match moving,” states Phil Brennan. “She was full scanned for modelling and textures to create as much of a CG version of her as we needed. All of the facial expressions had to be precisely match moved into the computer so that everything she does is replicated in there so you can add the second skin on top of her.” Martin Hill enjoyed the challenge. “Viper was fun. As well as obviously using lots of snake reference and shredded skins we looked at various other natural references. For the fresh peeling this included everything from gutted fish to lychees! We also made up two of our brave crew Martina Johansson and Emily Peters with a latex application and filmed peeling it off. This gave us a lot of great reference especially around the eyes.” Hill notes, “Human skin is always a challenge and our methods are constantly being refined. There are many factors involved in not falling into the trap of looking too plastic, too waxy or too consistent. Reference is always the key!”
Weta Digital was responsible for the ageing and de-ageing of the industrialist who seeks to become immortal at the expense of the Wolverine. “We had a young [Ken Yamamura] and old Yashida [Haruhiko Yamanouchi],” states Phil Brennan. There were various levels of prosthetics and make-up for each one. For each of the shots we generally shot on the younger side, then added in the age and took it away again for each shot. There were some particularly difficult shots where we go from one to the other. Those are the ones where you see the most change. It goes from being old to young Yashida and the other way around where he ages in the end.” Brennan explains, “Off to the side we had the older Yashida actor Har replicate Ken’s performance several times and we shot that with Red Epics from several angles simultaneously. We had him lit so it would match Ken’s lighting on-set; that gave us a whole bunch of textures and motion that would roughly match what one actor was doing to the other actor. We did that in a controlled environment. It would have been too difficult to do that on-set because it would take forever and slow the whole crew down. We did that off to the side as a separate green screen shoot. The effect was a combination of on-set photography, of lots of bits and pieces of this green screen shoot that we did with multiple cameras on the other actor and a whole bunch of CG to fill in the gaps.” Martin Hill remarks, “The ageing and de-ageing effect was very involved and essentially full 3D rebuilds of the characters perfectly match moved to the performance on the plate, so we could blend shape the model/textures separately and non-uniformly across the faces. The compositors created the blending mattes of the various layers in UV space to the timings so that the deformations, textures, and displacements all worked cohesively. There was also a lot of astonishingly complex compositing work to perfect the result; in certain shots we retained the teeth, or eyes from the plates.”
“The bear scene was done with an animatronic bear,” reveals Phil Brennan. “We got the footage back, cut the sequence and when we looked at it we were like, ‘You can tell it looks animatronic here and there.’ We broke down those areas and went back in match moving on top of the animatronic bear and replacing parts of it, the lower jaw and inside of the mouth. We were working with the animatronic performance but trying to enhance that performance and add in some more subtleties and nuances to give that extra level of realism.” Martin Hill remarks, “For the Bear, out textures supervisor Gino Acevedo did a photo shoot of a young grizzly in California which we aged and based our bear on. This was also helpful reference for the digital augmentation of the practical night bear.” A trademark ability of Wolverine was easy to underestimate. “The main issue was getting a tight match move on Hugh's hands. We had a lot of fun with the claws getting sliced, with the heat effect of the sword, and leaving the discoloured metal, similar to around arc welding.”
Previs and postvis was an indispensable tool. “Halon did a great job,” states Martin Hill. “The Bullet Train Sequence was shot on green screen very similarly to the previs on-set, which gave a good indication of the beats of the non shot elements. Silver Samurai was changed drastically for most of the sequence. The set was changed to be more vertical before shooting then we went into blocking from the plates. However, it gave us a good starting point for discussions for motion of the Samurai, and certain sections [Logan being pulled up by his claws] are very similar to the previs.” Hill observes, “There were many tough challenges on individual shots: Viper peeling, Yashida ageing, the practical bear augmentations and maintaining the consistency of the Bullet Train sequence for that many shots. Another challenge was giving the Samurai's performance the weight of his design, whilst keeping his motion nimble and acrobatic.”
“RSP had previously worked with the Studio 20th Century Fox and VFX Producer Jamie Stevenson [Charlie and the Chocolate Factory],” states Rising Sun Pictures Visual Effects Supervisor Tim Crosbie. “We had created a wide range of effects that suited the requirements of the project. Additionally the film was shot in Australia, as part of the Federal Government's Location Offset, and we were brought on as the largest Australian vendor.” Meetings occurred with Phil Brennan while the principle photography was taking place. “When I was on-set with him for the Japanese Ice Village Sequence we went through the initial briefs for what Phil needed us to achieve with Logan's claws, and the Ninja arrows and ropes. For the rest of the work on the show we had quite a few conversations using cineSync as the main creative tool to sketch in what was needed such as the Nuclear Explosion Shots, the Temple Fight and the digital extensions.” Crosbie remarks, “A lot of what we did had to sit with the surrounding shots so the creative brief was very clear from the outset. The Production Art Department did a beautiful job of providing us with reference and ideas that had been signed off by Jim so we had a great starting point. For the nuclear explosion Jim's wanted something that wasn't your typical library footage of a mushroom cloud and Phil did a fantastic job communicating what he and Jim needed creatively. Nick Pill [RSP Art Director] ran a few concept rounds on the explosion to get this into a place for Jim and Phil to get a good idea of what they wanted and then we spent a month working up our internal workflow processes so that we could iterate versions quickly for them.”
Opening the picture is a World War II scene that depicts the dropping of a nuclear bomb. “We shot a little bit of interactive wind but at the end of the day it was easier to do most of that stuff CG,’ states Phil Brennan. “We shot that sequence just outside of Sydney in an area that had a small beach transformed by the Art Department. We did go to Nagasaki and shoot a few plates but the majority of it was built and created in the computer. The explosion itself was looking at old footage of nuclear tests and of Nagasaki itself. There were a couple of photographs on the ground during the Nagasaki explosion which gave us some good reference to try to tie it in as much as possible with what happened. An actual nuclear shockwave moves at 600 miles an hour and if we had things move that quickly it would have been over in a few frames. We did have to take a little bit of a license and slow it down here and there so the scene could be more dramatic.” Crosbie notes, “We leveraged a lot of work that we'd done on previous shows, including the use of the bullet physics engine to destroy buildings using Houdini as the primary procedural application. We have a great team here who spend quite a bit of their spare time geeking out on the latest and greatest technology in this area so working on something of this scale was exactly the kind of challenge that pushes our team into top gear.”
“I'd like to introduce Prema Paetsch, our FX Lead,” remarks Tim Crosbie. “He put in so many hours setting up and then implementing the workflow we ended up using and the results were amazing.” Paetsch explains, “The mushroom cloud was achieved through a combination of techniques using Houdini. A highly art-directed smoke simulation with the required detail to sell the towering scale would simulate for way too long with uncertain outcome. We ended up driving a particle system through a basic smoke simulation for general motion and behaviour. This allowed us to intuitively re-shape, re-time and sculpt the basic particle cloud to work as a 'skeleton'. We then generated a library of independent and smaller smoke simulations for pyroclastic motion and detail. Finally, those simulations were then instanced onto the particle cloud with varying parameters to look like a cohesive mass without any repetition. Lighting and shading ended up being a comp-exercise with lots of control passes to achieve highly detailed shading and global illumination onto renderable volumes. The internal illumination was based on an old trick where thousands of little point lights were scattered within the volume and the attenuation distance tweaked until it looks right.”
“We used the Bullet engine within Houdini to destroy a detailed CG representation of the POW-camp, as well as buildings for the city of Nagasaki,” continues Paetsch. “The biggest challenge for the hero-destruction was a believable behaviour and choreography. We built a custom tool-set to glue sections of structure together and release glue on command to release fractured pieces in a staggered fashion. This had to fast and intuitive enough to allow for many iterations and flavours of destruction. Many, many layers of smoke simulations of varying density, look and behaviour were necessary to generate the shockwave as well as instanced debris, sand and dust particles, to fill up the frame with threatening chaos. The CG POW-camp proved useful again, as everything had to interact, collide and bounce off the geometry to look believable. Behind all this we had to generate a pyroclastic wave of very dense smoke, which shared some of the mushroom cloud development. The main focus here was even more close-to-camera detail and choreography, matching pyroclastic flow of volcanic eruptions.”
“We built the whole village in the parking lot by the Olympic Village outside of Sydney,” replies Phil Brennan when asked by the Japanese Ice Village Sequence. “There was lot of set extension work there.” A major challenge was encountered. “When Hugh is running through the village and he gets shot by arrows that have ropes attached to them. Hugh gets slowed down until he has about 40 to 50 arrows in him with the ninjas holding the other end. We got as much tracking information as we could. There were a few practical arrows sticking out of him but that was all done by Rising Sun; they’re particularly good with anything to do dynamics. They did a great job of getting the tension dynamics on the ropes pulling correct.” Tim Crosbie states, “Nick Pill [RSP Art Director] worked with Victor Gluschenko [RSP Senior Animator] to create a simple animated overlay over the top of the first cut of the sequence from Editorial. We ran through one or two iterations to get everything timed out right for Jim.”
“We built a simplified digi-double rig for Hugh from the cyberscan so that we could run a full match move of his body, legs and arms,” explains Tim Crosbie. “Our camera department had to make this as accurate as possible so that they could then place locators where the firing points from the bows were as well as the hit points on Logan's back which were all based on the original ‘pen and ink’ postviz overlay that Jim and Phil signed off earlier. This information gave our animators a great head-start; it's relatively simple to brief someone by saying, ‘Please animate the arrow's path from locator A to locator B, over this duration.’” The nHair plugin in May was used as the primary tool. “The ropes were a relatively simple procedural process once the animated paths were in place. They connected to the nock of the arrow and the original source point at the associated Ninja and we let the software run the initial simulation, after which we did jump in and modify a little where needed to get the best look for framing.” Crosbie states, “The rest was essentially a process of making sure that what we did was seamless with the surrounding cutaways. In many shots Hugh was wearing prosthetic metal claws while he did his own stunts. In the shots where the stunts were just too dangerous to have full length metal claws flying around we went in and added CG claws which obviously had to match perfectly to the lighting for each shot.”
“We worked on other sequences such as the Temple/Funeral Fight Sequence where we added Logan's claws for the more dangerous stunt-work as well as the Tokyo skyline when we shot off the location [in Sydney's Chinese Garden],” states Tim Crosbie. “Also, the LS Sequence, where Logan and Shingen square off was an interesting challenge; there were a few shots where we had to significantly modify the plate to help make the tension of holding Shingen's sword in Logan feel more real. One of our digital matte paint extensions was the one in the Yukon where Logan is coming into town after having to kill the Grizzly Bear. This was shot in Picton near Sydney in the summer and we placed mountains into the deep background as well as the street leading away from us to connect everything together."
“We did a lot of research online to get as much reference as possible so that we could do justice to the locations that needed to be seen on the big screen,” says Tim Crosbie. “Nick Pill spent a lot of time pulling all this together so that we could present concepts for discussion with Phil. As soon as we knew what we needed to hit, the work flowed from there. Nailing down a tight brief at the beginning means we could give our clients the best bang for their buck. After that it's just time, achieving the right level of detail for something to hold up on the big screen can take up to a month of an artist's time painstakingly ensuring that every detail is given the right amount of DMP love. Once we were happy with the matte paintings our compositing department picked up the assets to run the final integration into the shots.”
“The claws were a relatively simple process for us to build. I should add though that much of this is due to our Asset Supervisor Dennis Jones,” states Crosbie. “He's spent the last year or so building up not only great team but a really good pipeline for our asset builds. The workflow was pretty traditional, we pulled in an example plate with existing practical claws and the associated HDRI, from this we set up a global illumination lighting setup and placed out CG claws, as a turntable, into the shot so that we were looking at apples vs apples. When our claws are indistinguishable from the real thing then we know we're ready to start pulling these into the shots.”
“Other than Wolverine's claws [which, in addition to claws, also occasionally was an entire hand replacement], there wasn't a lot of augmentation & digital doubles for RSP's work on this show,” remarks Tim Crosbie. “One of our mantras at RSP is, ‘If you can see what we've done, then we're not quite finished yet,’” observes Tim Crosbie. “Although I guess that a nuclear explosion is a little hard to miss. I hope that all of our claw and arrow work might be called invisible effects; some of the digital extensions are certainly in that category. I'd like to think that if we showed you all of the shots we worked on you'd be surprised at how much real estate within the shots is pure digital.” Crosbie notes, “I can speak for pretty much everyone in the company when I say that working on stories in the Marvel universe, and Wolverine in particular, is about as much fun as anyone could have in our field.”
“Iloura was introduced to the project at the earliest of stages bidding a number of sequences during the pre-production period,” states Iloura Visual Effects Supervisor Glenn Melenhorst. “Sequences were awarded based on specific needs from the studio and VFX supervisory team.” The computer proved an effective communication tool with Phil Brennan. “We started the creative process by discussing each scene via Skype. From that point we evolved a range of look development and concept art for the healing shots and for the pin bed sequence. Smaller shots, such as the shot where Wolverine pushes a piece of glass through his cheek, were more practically handled – we took the lead and discussed creative merits of the work once in progress.” Melenhorst remarks, “We dealt mainly with environments and objects, like the pin bed. The only sequence that dealt with matters of character was the burning and glass wound healing shots; for these we had discussions around how the Wolverine healed himself. We handled Viper’s tongue and poison FX but they too didn’t require much more than a phone call and a couple of still frame paint overs to get started.”
A cinematic moment which dramatically demonstrates the healing ability of the Wolverine occurs in a well during the Nagasaki nuclear blast. “Our prosthetic guys did a great job of making Hugh look burned,” states Phil Brennan. “But the shots going from him being burned to healed had to be all CG. The live-action part is him when he is fine. We digitally applied all of the burns to him and digitally took them away revealing his pristine skin.” Glenn Melenhorst explains, “Our first challenge with this shot was the soft tracking of Wolverine himself. James Mangold’s preferred take of Wolverine standing up after the blast was a take without any tracking markers. As a consequence Iloura Artist’s Paul Buckley and Sam Jensen hand tracked freckles, clumps of hair and patches of skin, and painted scores of custom blend shapes to make our digital sculpt of the Wolverine stick. The process took the pair of them over a month to complete. After that process we made a start on the actual effect. Drew Wood Davies, our lighting lead on the feature, reverse engineered the lighting form the environment for us to relight our digital double accurately.” Melenhorst observes, “The effect would have looked like a morph or a dissolve if not handled carefully. One of the most important aspects of the effect was the multi-layered stages the healing went through; from displaced, charred skin through to wounded scarred skin, to pink, new skin and then healed. The shot is a fully digital asset until the healed skin where we basically revealed the original plate. Coupled with that was our hair system, which grew strands that ultimately were groomed to match the Wolverine’s hair on-set. We staggered the effect so it did not happen all at once, by selectively growing hair, and again with offsets. The effect forces your eye to jump from healing patch to healing patch and before you are aware of it, it’s over.”
“Originally, the pin bed was to be mostly covered as a practical set piece but as things happen on a live set, the decision was made at the 11th hour to film a lot to the shots on the green screen prop,” states Glenn Melenhorst. “We had very good reference for the pin bed as the real bed was used in a handful of shots. What differed between the real and green screen bed was where the actor was positioned. Having no pins to reference to, he was slightly diagonal on the set piece and when our rigidly straight pins were added the line-up was quite untidy, with the actor cutting across lines of the pins. From that point on we along with Phil Brennan art directed each shot down to the individual pin head and their reflections to make each shot work.” Melenhorst remarks, “Other than the healing Logan, the pin bed and nanotech beetle, we finished the skyline extensions and set extensions for Yashida’s apartment, FX for the bar fight including the glass in the cheek shots, Viper’s tongue and poison FX and numerous other smaller bits and pieces.” The visual effects work was not meant to go unnoticed. “Apart from the set extension work of the Tokyo skyline, and our 3D rain and ocean effects in that particular scene, all of our wok was very visible – for example, healing wounds, the snake tongue and pin bed. It would be a shame if there were invisible!”
“Method came in to the project during production when specific sequences had grown and additional work was required,” states Method Studios Visual Effects Supervisor James Rogers. “Phil had a clear idea of what he wanted in each scene, so we had very direct briefs for a lot of our shots. Some shots, like any production, had an organic development which changed as we went further into them – particularly with multi-vendor sequences.” Rogers remarks, “For our shots, we were mostly concerned with making a believable environment for the characters to exist in. It's not so much about manipulating the character as it is manipulating their circumstances – where are they and why? How does the light around them look, how can we bed them completely and convincingly into a scene [from a green screen or studio shot]? Dealing with an alternate universe like this requires careful attention to making something look real and tangible, but has certain aspects which make allowances for the heightened reality of the story.”
“Method Sydney did environments for a Yukon town, Tokyo, including downtown scenes, the bullet train interiors, Shinjuku station, Noburo's [Brian Tee] apartment, and a variety of claws, eyes and billboards,” states James Rogers. “The Tokyo sequences were challenging as the original shoot elements and references source were shot on public streets. While they looked busy, it was a challenge to make the action work for the purposes of the film, especially in terms of continuity. We used a lot of techniques to patch the backgrounds together, in addition to perspective tricks to make them work for the studio-shot foregrounds. The longer sequence of Yukio [Rila Fukushima] and Logan driving through a town in the Yukon also required similar manipulation to bring a consistency to the backgrounds and the rain.”
“Regardless of the film, the practical effects give us a reference for how we will treat and composite the digital effects,” remarks James Rogers. “It is an invaluable help to have as much practical effects in a scene as possible, as they give you some absolutes and inform the way that you will build the shot. Generally we will request as much information as possible – edits, stills, on-set pictures, as well as the shot-specific footage. The more we have, the faster we can pull a look together.” Rogers states, “We did a lot of Japanese scenes. Most of the material we were dealing with had the actors shot on green screen or partial sets, and we extended them, or built the backgrounds from reference footage that was shot in Japan. It was an interesting project for me, as my partner is Japanese, and I've spent a lot of time in Tokyo. A lot of the locations were very familiar to me. I think there is a certain style to the way Japanese streets look, which was fun to retrofit into certain plates shot in downtown Sydney. Little things, such as road markings and signs are subtly different.” There was not much in the way of CG augmentation and digital doubles. “Not for the shots we worked on, but we did use some large crowds to fill in the backgrounds in some of the Japanese scenes.” Unlike its sister facility Iloura, the contributions of Method Sydney were intended to seamlessly blend into the background. “Being that most of the shots we were worked on were environment shots it was our focus to make this work invisible.” Also recruited for The Wolverine was Method LA which did some fixes on the Easter Egg Scene at the ended that serves as tease for X-Men: Days of Future Past (2014).
“We've developed a great relationship with Fox and Joe Conmy by creating visual effects for Die Hard 5 ; Joe was kind enough to introduce us to the team on The Wolverine and get the ball rolling,” states shade vfx founder and owner Bryan Godwin who served as a visual effects supervisor on the project. “Phil is a fantastic supervisor. We had a great time working together. Phil is a perfectionist but is unique in his ability to communicate what he's looking for very specifically. Every iteration pushed the work forward and made the shot better. We communicated occasionally via cineSync, but here at shade we had the advantage of being in Los Angeles, so that Phil could drop by any time and talk through the work in person in our screening room.” Godwin remarks, “We did a little over 110 shots for the film, and the bulk of them had CG components or matte paintings, but of course there is always the odd tracking mark or rig removal that needs to be taken care of by the paint team.” Being a boutique facility was an asset when it came to managing last minute tasks. “We did pick up a couple of last minute sequences due to some reshoots and editorial changes. The best thing about being a shop of our size is that we are nimble, flexible, and able to accommodate last minute shots right up to delivery. Shade has managed to assemble the perfect team of experts without having a ton of overhead, and that allows us to really be effective in these situations. By having a smaller team of more senior artists, we can tackle a lot of shots very quickly.”
Shade VFX handled the Love Hotel Scene. “When Jean Grey was communicating with Logan in this sequence, what seems like a simple composite was actually a precise creative exercise. It took several iterations to find a look for Grey's ethereal presence that was tasteful and filmic but communicated the surrealism of the moment.” Background replacements had to be integrated into the shot. “We created a fully 3D set extension for the exterior of the Love Hotel to replace a downtown Sydney location. In addition to creating the CG extension we added Japanese signage and street markings to convert Sydney into a convincing downtown Tokyo. We also created and added a Tokyo skyline to the final airport sequence for continuity.” Godwin remarks, “We were responsible for adding CG claws in a number of shots. Additionally, we added CG wounds and cuts to Logan's death sequence in the laboratory when he regains his powers.”
“We were involved thanks to recommendation from Amber Kirsch, as she was happy with UPP work on Die Hard 5 for Fox studio,” states UPP Visual Effects Supervisor Viktor Muller. “Phil was great and accurate in his description what he is looking for. We got as well very simple but very helpful previs from the editorial.” There was a sense of collaboration with the production team. “We were very happy because we got great chance to be really involved into creative process. They like our ideas and from other side we received good comments.” The Prague based facility is known for producing CG environments. “I hope that next time we will have a chance show to Phil that UPP has many more skills. The true is that we have big reference for environments and full CG shot from many shows. This experience helps to avoid mistakes and safe some time during the process.” Vegetation needed to be researched for the establishing shot of the Yashida cottage. “For us was probably the most important to find out what plants we have to create and how far we can go with physical effects and atmosphere. We became part of the show on the very end, so we were little bit afraid about continuity with other shots.” A particular aspect proved to be difficult. “One of our challenges was a stormy ocean. There was not enough time to create CG stormy water. But we solved that easily as I took camera and shoot the elements on a real ocean in the night.” Muller remarks, “We created a lot of layers for rain, wind, atmosphere, interactions and reflections which were use just with very small opacity. This is little bit of secret how to create good looking full CG shot. Some plates were really amazing, so shame that you can hardly see them!”
“Several years ago I worked on Knight and Day with Jim who called me to work on the previz for The Wolverine as soon as it looked like he would land the deal,” recalls Halon Entertainment Previs Director Clint Reagan. “Jim used the previs to explore his ideas and to get a feel for opportunities in the spaces he had yet to see in reality. Jim usually spoke more to the feel of the action and the scene cut together focusing on the intensity and clarity of the shots in terms of the idea being portrayed. James was always looking for what kind of mood was being created from the variety of action. Michael McCusker [The Amazing Spider-Man] being the editor would get my cut and then sit with me to recommend shots, timings and trims that would help the cut to read more clearly. Michael and I also found a way to work back and forth between his and my editorial packages so that he had my initial assemblies for his initial pass of each sequence and I could revise by working from his cuts. Phil and I often sat to review the cuts and check the viability of the ideas for VFX before presenting to Jim. This way I could prepare for how he might approach the work for finals and keep away from any aspects that would prove more difficult to accomplish than would be worth the effort.” Reagan remarks, “If anything we used the previous Wolverine feature as a comparison for how grand and full of spectacle Jim did not want to take this film; he often mentioned wanting a more personal movie with more personal action closer to the hero than in the previous film.”
A close collaboration was developed with the stunt team. “Working with David Leitch [Jupiter Ascending] was great,” states Clint Reagan. “He and his team of stunt artists created so much exciting action with such intensity that replacing all of it with previs would take too long to make it worth even trying to match the intensity he already had. So Jim had us cut previz in and around Davids stunt-viz. I would add some full previz shots and often use 3D and 2D tracking of his stunt-viz footage to add backgrounds and characters to his amazing performances. It worked really well focusing on each of our mediums strengths, his intensity and realism from performers (it was real after all) combined with his frenetic shooting style and my ability to contextualize any shot and create wider scope shots to make the overall edit work well to clearly tell the story that Jim was liking.”
“As far as Train vs. Silver Samurai Sequences these were approached quite differently from one another in previz,” remarks Clint Reagan. “The Bullet Train Scene was generated and explored quickly with one of my four teams of artists doing various sequences. The Train Sequence was built quicker and more straight ahead as an exploration of the speeds being travelled. We worked stream of thought will little time to dwell on the structure of the scene. I think that helped keep it an anxious scene in previz and on screen. Whereas the Silver Samurai Fight was planned many times in many ways and situations. There was a stunt-viz pass of portions of the fight from David Leitch, and a separate first previz pass in a cave location. We did a motion capture shoot of many portions of the fight, capturing David’s talents performers doing some amazing martial arts choreography and flips and attacks. Many of the main ideas were found at this phase. Then we came back to the sequence later on to move it to the laboratory tower environment and conform or change ideas to fit the new situation. Much of the motion-capture was lost in this process as the performance would change so much that it was cleaner and better looking to delete the motion-capture and proceed with key-framed animation. In spite of that the motion-capture performances still started the exploration of the moves and moments that could be created. So the Samurai Fight was explored more thoroughly and the Bullet Train as a high speed stream of thought approach that stuck without needing more exploration.” Reagan reveals, “The challenge for me with the Bullet Train Fight was the speed at which we had to deliver the full sequence to stay on schedule. I wanted to explore more ideas but had to work that sequence at the speed of thought and get it cut in fast. In a way I think that gut reaction approach added to it a lot.”
Mapping out the scene where the nuclear bomb is dropped on Nagasaki required some intervention from James Mangold. “I was so stuck on the research I did for how high the airplanes were when they dropped the bomb and how it would affect the shots. The planes were high enough that they would not have been visible from any shot on the ground. Jim’s reactions to me did a lot to break me of the last vestiges of adherence to physics and realism in movie making, which my work has benefited from ever since. In addition, the Nagasaki section that we prevised although different I think than what appeared in the film had a lot of time spent on building the progression of the effect for the blast. This was tackled by Earl Hibbert the talented artist who spent a lot of his focus and valuable time adding each layer to the idea till we really felt like we had created something a bit different than what you might expect to see from an atomic blast. That was Jim’s goal to make it a bit different."
“We were asked for some shots exploring the idea of Viper peeling her skin,” recalls Clint Reagan. “My artist who did that did a great job building a quick rig that would peel her apart, but the shots I got for her peeling her skin off were not appropriate for even a PG-13 film! I was ribbed for those shots for months.” Reagan states, “I enjoyed the Bear Sequences we created for the film early on. It was much bigger and physical than what ended up being shot. We created a very intimidating bear hunt and fight with Logan that really would have taken the breath away at full screen. There was a challenging night when a note came down and needed to be turned around fast to fix a scene with a bear in the movie. The VFX producer at the time, Greg Steele and I stayed up quite late [or early since we hit early morning] and slammed together a pretty exciting scene with the bear stalking Logan in the forest leading up to a mountain side tumble. It turned out really well for as fast as we put it together. In the end it was too much for something that didn't strengthen the story enough compared to the effort needed to create it. But that’s where exploring ideas in previz can be so valuable to production, we can help figure out what should be in or out before you have to film it or commit the Art Department and VFX talent on it.”
Major fight sequences had to be orchestrated. “The challenge with the Funeral Fight was taking the great stunt-viz that David Leitch and crew shot and then fitting it into the actual location in Sydney that we had built to scale in previz,” notes Clint Reagan. “Giving the scene a coherent flow and story that worked with the location where the shoot would be was a challenge but honestly that is the fun part to me, figuring out how to use what you have to get the story made well.” Reagan remarks, “The village fights challenges were in using a limited space to accomplish all the needed shots without creating the feeling that we were on a limited space. So we reused portions of the set that I had from François, the production designer, to make the streets feel longer than they actually were planned to be.” The continuous production delays caused further complications. “The biggest challenge was that the show was off and on over a year plus with 2-3 month pushes and then delays in schedule. So each time I started over with a different team of artists due to availability. But in spite of that we managed it well and kept adding to the work in helpful ways.”
“It’s a great fight sequence,” replies Phil Brennan when asked about the violent mayhem that erupts at the Temple. “A huge amount of credit has to go to Dave Leitch our stunt coordinator and also our second unit director. We did previs on that. We had stunt-vis and storyboards on that going way back. It took us a long time of planning but they did an amazing job of the choreography of the fight in that sequence. That’s a sequence which is stunt based and we had all sorts of claws effects and adding Tokyo in the background, changing some of the pieces of architecture and lots of rig removals but nothing particularly complicated effects wise. A lot of the credit goes down to the great stunt work and the fight choreography. We’re there to provide the finishing touches to things. To clean up where there had to be crane in the shot or a stunt rig. All day-to-day stuff for us in the visual effects world but also helps to bring the sequence together. Things do get crazy we have to do something that a stunt person can’t do and we work with a digital double. We didn’t have any of that in the Temple Fight Sequence. We had at the bullet train and a few other places.” As for seamlessly integrating digital doubles careful match moving and accurate lighting are critical. “Usually you make that takeover point happen when there is a ton of motion blur or the camera is moving violently so you can disguise the point of the takeover from live-action to CG.”
A number of creative issues needed to be address. “There were three or four sequences that were technically challenging,” notes Phil Brennan. “The explosion at the beginning, getting that camp destroyed and the explosion in the background. The Bullet Train Sequence was something we spent a lot of time on and the samurai at the end was a huge number of shots. The Viper [skin peeling], and the ageing and de-ageing shots even though they’re little moments in the film were extremely difficult.” Some minor adjustments were made to accommodate the 3D post-conversion. “A lot of that work was done by Stereo D. We did have a separate stereo supervisor who handled all of the conversion work so he did all of the communication with Stereo D. From our point of view we had to be conscious of it. Sometimes when things are turned into 3D it brings up something you might have never thought about before. Suddenly things can become a feature in a shot that you never expected.” Brennan, remarks, “Hugh was a joy to work with for everyone but with visuals effects especially he was so easy to work with. The other actors were all fine too. People expect there to be a lot more visual effects and they understand that we sometimes ask for weird things.”
Productions stills and visual effects images courtesy of Twentieth Century Fox Film Corporation, Weta Digital, Rising Sun Pictures, Iloura, Method Studios and Halon Entertainment.
X-Men Character Likenesses TM & © 2013 Marvel Characters, Inc. All rights reserved. TM and © 2012 Twentieth Century Fox Film Corporation. All rights reserved. Not for sale or duplication.
Many thanks to Phil Brennan, Martin Hill, Tim Crosbie, Glenn Melenhorst, James Rogers, Bryan Godwin, Viktor Muller, Prema Paetsch and Clint Reagan for taking the time to be interviewed.
Make sure to visit the official websites for The Wolverine, Weta Digital, Rising Sun Pictures, Iloura, Method Studios, shade vfx, UPP, and Halon Entertainment.
The Wolverine Interviews - Rila Fukushima
The Wolverine Interviews - Tao Okamoto
The Wolverine Interviews - Will Yun Lee
The Wolverine Interviews - Tao Okamoto
The Wolverine Interviews - Will Yun Lee
Trevor Hogg is a freelance video editor and writer who currently resides in Canada.