Grimm Reality: The Making of Snow White and the Huntsman
Wednesday, 18 July 2012
Trevor Hogg chats with visual effects supervisors Cedric Nicolas-Troyan, Phil Brennan, Todd Shifflett, Bryan Hirota, John Moffatt, Nicolas Hernandez, Edson Williams, Steve Moncur and Angela Barson as well as effects supervisor Lindsay MacGowan, Nvizage owner / founder Martin Chamney and creative director Henry Hobson about their work on Snow White and the Huntsman...
“I have a long standing relationship with Rupert [Sanders]; for the last eight years I did a lot of his commercials so when Snow White and the Huntsman [2012] came along he wanted me on the job,” recalls Visual Effects Supervisor Cedric Nicolas-Troyan (Solstice) as to how he came to get involved with the re-imagining of the classic Brothers Grimm fairy tale. “At the time I had retired from visual effects and was directing commercials on my own, but the opportunity was too big to refuse.” The two veteran collaborators share a similar creative style and sensibility. “It was more about figuring things out together and less about me trying to decipher what he wanted.” Having to deal with a director helming his feature film debut was not a major issue. “Ru is a very smart guy who is very collaborative and doesn't get stuck with the idea of his way is the only way; he listens, processes and always tries to make the best educated decision.”
Unlike most films two rather then one visual effects supervisors where responsible for the project. “It was a bit of a leap for the studio to jump ahead with a first time director and visual effects supervisor on a [movie of this] scale that has a very short turnaround in post,” admits Cedric Nicolas-Troyan. “Having a second visual effects supervisor helped everybody to sleep better at night but very soon it became something else. It was great to work with Phil [Brennan]. We have the same type of background and approach to the work. I had to design a lot of the creatures that you see in the movie plus work with Ru on a daily basis to develop the look of the film. Later I directed a few 2nd unit scenes so I was not on-set for the main unit. It was amazing to have Phil with me; otherwise, I wouldn't have been able to do that at all. In post because the turnaround was short we were able to divide the work. Phil went back to the UK many times to look over our UK vendors while I was staying under the California sun.”
“We wanted to keep the fairy tale aspect somewhat grounded,’ states Cedric Nicolas-Troyan. “The first thing was not trying to explain everything and anything. The second thing was to avoid any type of light, glowy or over-the-top effectsy FX as much as possible. We tried to root any type of fantasy aspect so it is not realistic but still feels real in the world we created. We also tried to keep the design unified and away from a patchy all over the place type of feeling, especially, when you have to cover so many different types of visual effects.” Phil Brennan [Terminator Salvation] notes, “Even though there is the underlining fairy tale theme to it, it’s a gritty adventure action film. A lot of that comes from Rupert. Rupert didn’t want things to look over-the-top; he wanted things to have a gritty dark believability about them so we were always leaning in that direction.” Nicolas-Troyan observes, “It's funny how everybody imagines Snow White as the Disney character. The original tale is much more vague and dark.” Brennan sees more of a connection with the fantasy trilogy helmed by Peter Jackson than the medieval story directed by Ridley Scott. “The Lord of the Rings [2001 to 2003] comes in because there’s a lot of scale. There’s a lot of journeying so it has that same epic feel to it. Robin Hood [2010] I’m not so sure. We’re drawing on anything that has tons of battles.” As for what will make this version of Snow White distinct, he remarks, “That comes from Rupert himself; he has a unique vision and visual style so bringing that into play, at least visually, gives the movie a unique look.”
“Due to the fact that we shot a lot of the film at Pinewood Studios on stage and in the parks adjacent, there were a lot of extensions of villages, castles, vistas, and interiors,” states Cedric Nicolas Troyan. “I don't think there are a lot of shots left in the film where we did not add something.” The computer trickery was not confined to the magical elements. “Beside the Troll and the Wood Fairies, which are obviously fantasy creatures, every single animal such as the birds as well as the insects in the movie are entirely CGI.” Practical effects were also utilized. “It’s Rupert’s, Cedric’s and my style to always start with something practical,” explains Phil Brennan as to the approach taken to build the various cinematic environments. “The Enchanted Forest being the most obvious one where it’s hugely enhanced; we always made sure that we had a good in camera base. When you’re going for a gritty realism it’s hard to make an entirely synthetic environment look good with the rest of the film.
“The most challenging [aspect] was the dwarfs because there are eight of them during half of the movie,” reveals Cedric Nicolas-Troyan who had to incorporate the iconic fairy tale characters into various scenes where they have dialogue, dance, fight and run. “We had a limited amount of time with them so any VFX invasive technique was completely out of the question.” A group of little people devised by legendary fantasy author J.R.R. Tolkien loomed over the desire to create something that had never been done before. “A Hobbit is a fully proportionate person scaled down whereas a dwarf is a normal person with limbs that did not develop which is a very different thing. We had to make the audience believe that those renown actors, who we have seen in many other movies, are actually dwarfs with short limbs and a lower center of gravity which makes them move sideways instead of forward backward. We had to use a lot of different tricks to achieve that but in the end I think the audience will believe.”
“When you first read the script that’s the most intimidating thing you come across,” recalls Phil Brennan. “‘Oh, my God, how do we turn these big guys into dwarfs?’ If you have unlimited time and money there are all sorts of ways that you can go about it.” Time and budget constraints could not be ignored. “We had to work out ways where we could do a lot of the work in camera with some fairly old school cheats, being careful with lens choices, angles and the relative heights of things. It involved working closely with the wardrobe department to make sure the costumes simulated the different proportions to help make the actors’ arms and legs look shorter. We had a great prosthetics guy named Dave White who gave the dwarf-size doubles the actor’s face; that only holds up for a certain degree and held up better on some people than others. We did a lot of visual effects replacement work 2D and 3D. We did some shrinking of things where we shot each actor individually within a shot.” Brennan continues, “A lot of the re-proportioning stuff was handled by Rhythm & Hues and a lot of the face placement type work was handled by Lola. The conventional way of doing face replacement is to have the actor there on the day of shoot and to shoot their face in the same lighting so to get face replacement element. When you have eight actors who are moving around fighting and doing all that kind of stuff it’s not feasible to do that all day. Production wise you can’t afford to spend an extra two or three hours on a shot to get all of the positions of all of the faces for eight actors. We ended up doing a lot of that stuff later using Lola’s face replacement rig.”
“As the script evolved the visual effects evolved with it,” remarks Phil Brennan on the production which required 1300 visual effects shots. “Because of time and budget we had to take the divide and conquer approach based on the design of the film,” states Cedric Nicolas-Troyan when explaining how he and his colleague select the various VFX vendors and achieved a unified look throughout the picture. “Animals, creatures and all ‘life’ by design went to Rhythm & Hues. The dark, cold and ‘death’ by design went to Double Negative. The battle and soldiers went to Pixomondo. The Mirror Man went to The Mill. We chose the vendors based on what we believed were their forte.” Bringing the Enchanted Forest to big screen was the responsibility of the main VFX vendor for the fantasy adventure. “Rhythm & Hues has a great pipeline for creatures and we have a lot of creatures in the film that have to look photo-real and be fairly close-up,” states Brennan. “Pixomondo did the opening battle where you can see from the trailers the shattering soldiers; they also did a lot of generic battle stuff in the end battle – pyro duplication, explosions and fireballs.” The pivotal character of the Mirror Man was given to a British boutique VFX facility. “Both Cedric and I are familiar with The Mill. It is a fairly small but important part of the movie.” A specific look needed to be developed. “We didn’t want it to be the Terminator going way back to T2 [1991].” A solution was to add an additional element. “The option of metal cloth instead of liquid gave us that difference,” explains Nicolas-Troyan. “It gave us also the elegance that I was looking for.” Another British VFX company was responsible for a scene involving some poisonous hallucinatory mushroom spores and Snow White. “It’s a scary trippy sequence in the Dark Forest,” remarks Brennan. “Double Negative handled that part and the Dark Faeries at the end of the film which are made up of razor-sharp shards of obsidian and can reform into different shapes.” Also helping out on the $170 million Hollywood production were Lola VFX, BlueBolt, Baseblack, Legacy Effects and Nvizage.
“Cedric has this wild extreme personality about him and Phil is more internal,” observes Rhythm & Hues Visual Effects Supervisor Todd Shifflett. “They were both clear on what they wanted to have done. There were certain sections that Cedric was much more tuned into from the get go; he always referred to the Troll as his baby.” Dealing with the two supervisors was not an issue; however, the tight production deadline was problematic. “Because of the schedule we needed a lot of answers quickly for, ‘What types of animals are we going to make? What kinds of plants do you want us to make? What colours should things be?’ All of these questions need to be answered early on in the visual effects spectrum and Rupert did not have a lot of time deal with because he had a whole film to be putting together. Unfortunately, for the filmmaker by the time he gets filming wrapped up and is in a position to put a lot more of his focus onto some of the visual effects, a lot of the work is too late to go back to change.”
“The dwarfs were certainly a more unusual challenge than some of the traditional visual effects like inserting a CG character which has become common place,” states Todd Shifflett. “At the beginning of the project they were scheduled to be a much larger portion of the visual effects budget. We originally thought we were going to have a hundred more shots of dwarfs than we did. Most of the filming tricks we did on-set in order to try to cheat the look, and the in-camera effects worked so well we didn’t have to go back in to try to do some visual effects later in post.” A lot of research development went into devising 3D and 2D solutions to emphasize the disproportions of the limbs of a dwarf compared to an average human being. “It was important to the filmmakers that we maintained the difference in proportions rather than simply take the actors as they normally are and shrink them to make a tiny version of themselves. The disproportionate nature in the shrinking makes it tricky. You’ll have something that seems to work well but then suddenly the actor will put his arm out to the side or he’ll swing an axe around or he’ll drop something onto the ground and it changes everything. If it’s a sword we didn’t suddenly want that to get longer or shorter. There are all of these little confusing things that might happen on an individual perspective and there were eight different dwarfs in the shot. They’re rolling long takes. They’re doing different things each time. After six or seven takes they’ll turn to us and ask, ‘Did you get what you needed to do the effect?’ We’re thinking, ‘Somewhere in there everyone at one point did something that will work.’ You’re hoping that it all works out because there is so much overlap you can’t keep up with it.”
“You learn quickly that you don’t have control over a lot of things,” admits Todd Shifflett when discussing how he was able to keep track of all the complexities involved with the dwarfs effect. “It’s not within the scope of a human being brain of being able to try to keep all of this together. ‘I remember from these 15 different takes if we take this one for this actor and this one for the other actor it will suddenly piece together and it will all look like they were working.’ That’s what editorial is for. There was a lot of going through and working with editorial. A number of times they would pick takes they liked for the action and we would go back to them and say, ‘Unfortunately, although it is nice for the action it doesn’t work for us to deal with what’s needed to pull the effect off.’ Editorial was accommodating and understood that was part of the deal in going through with all of this; they would go back and would find something else that did fit the criteria.” The filmmaking process was hectic. “We didn’t necessarily know what job we were going to do on which day and suddenly be called up and say, ‘By the way, right now we’re going to film one of these shots which we’ll need motion-control for.’ We had no idea of what the shot was even going to be 30 seconds before and so it’s a matter of, ‘Okay, now we’re going to have a motion-control shot. What do we need to have these actors do? What path do they have to walk down? What do we have to make sure their timing is? We’re going to do a couple of different takes for each actor and run each one separately to make sure they’re not going to overlap one another. The dichotomy of trying to pull that together on the moment and finally seeing the end product all in one piece always startles me. I look at that and think, ‘Oh, my God. Look at that. It looks like they’re all there together.’ And that was never the case!”
Rhythm & Hues was handed the task of making the Enchanted Forest a cinematic reality. “From my standpoint, they ended up dialling back a lot of that,” remarks Todd Shifflett. “I wish we would have been able to go farther down the road of adding more of the plant life we had originally planned to put in there. Unfortunately, probably do to again the shorten production schedule where we’re building plants that the director eventually decides, ‘That’s not what I want in my forest.’” Rupert Sanders’ concept of enchantment was not of the grandiose variety. “What it turned out to mean was much more about adding lush greenery, moss to the trees, and splashes of colour here and there throughout the forest. Whereas when we went into the project we thought, ‘This could mean anything. We don’t know what it’s going to need. We need to be fully prepared to blank the earth with plants and foliage and put them anywhere we want to.’ There was a lot of research and development that went into being able to take different species of plants on different parts of topography which we did use occasionally.” Shifflett continues, “We wanted to try and work with different methods of gathering the imagery. How can we turn it into a 3D set of geometry so we can then put something on it? Or what methods would we use in order to take various 2 or 2½D elements into a scene and incorporate some motion off of them? For example, we might have moss that would be hanging from the trees. If the wind is blowing there needs to be a little bit of movement onto the foliage. We would take our painting that was being rigged off of a tree limb and we would run it through a cloth simulation to get some billowing nature integrated into the plant. Or we would place foxglove out into the field, these long stemmed flowering plants; we needed to get some motion onto them to make them feel like they’re actually in that world.”
“We’ve done a lot of creature work,” states Todd Shifflett. “It’s always being improved upon.” Background animals had to be altered for the foreground treatment. “All of a sudden we’d find out that our director wants to punch in and do a close-up. ‘Oh, we don’t’ have that stuff.’ Now we have to take this creature and put him back through our pipeline of getting our textures up in resolution, and a better groom on them.” Shifflett mentions a couple of examples. “The rabbit stood out to me because he was one that was never meant to be a full screen animal; he ended up getting a couple of shots. ‘Lets punch in on him and have a faerie sitting on his back.’ There are a couple of shots of the badger where we punch in closely onto him. The badger was originally suppose to be some background character that you might happen to see wandering around in the bushes; we did a good job going back in and giving him more detail.”
“A Western European Raphaelite kid was the ideal image for Rupert in regards to what the Good Faeries looked like,” explains Todd Shifflett. “He had chosen two children to play the part of the faeries and generate facial motion-capture for us. We modeled our faeries after those two children and tried to use that facial motion-capture wherever we could. Unfortunately at the time there was no scripted section for the faeries in the film so they were getting a random action that doesn’t work so well when you want to go back later on and put something specific in. You need to work in a certain direction if you’re going to use motion-capture.” To compensate digital keyframe animation was utilized. “We used that motion-capture as a reference for our animators to be able to go in and see the real details of the organic nature of the animation. These tiny details might be a flinch of the eye or a quick firing muscle that doesn’t seem like to do anything in particular; all of these little things add up to feeling of a naturalistic movement, whereas if we just have a lot of nice blend shades it might provide nice animation from an animated movie situation but it wouldn’t necessarily provide an organic natural animation. What we wound up doing is driving our animation by hand and going back referring to sections and go, ‘Look at what’s happening here. We need to get this feeling of the cheek movement happening in our own hand driven animation.
“The Troll concept was originally done by Cedric,” states Todd Shifflett. “We created a digital sculpture and built the troll from the ground up. What was nice about the troll was that we had a lot of backstory; there was a whole world that Cedric had come up with for why this troll exists and what kind of creature it is. It’s not necessarily an evil creature; it happens to be out there in the world. It hides in these places that have rocks and tree stumps, and have over the centuries become where civilization have built their bridges; that’s where we get the Troll Bridge. Any time we doing something with the design whether it was, ‘What are we going to do for his knuckles or his fingernails?’ We could go back to that backstory and understand that he doesn’t walk upright. He’s digging with his hands in the ground a lot so we give him a lot of earth onto the hands. The Troll doesn’t use his sight so much because his eyesight is poor so what he’s mostly seeing is mostly a blur; he needs to see some movement to be able to pick up on it. That worked well for us to try to understand while Kristen is standing still what does that mean for him in terms of being able to assess who this person is. ‘Does he smell her? What are the senses?’ The more backstory you have the better it becomes for the animators to be able to work.”
“There is the Great Father; a stag that is a life force in the forest,” says Todd Shifflett. “Yet another creature added that Snow White would have to interact with. Some of the interesting things with that was very late in production it was decided, ‘By the way, this guy shouldn’t die. When he gets shot with an arrow he should burst into a butterfly.’ It’s like, ‘Wow! That’s a big note.’ Those little surprises were interesting to deal with on the project. ‘Wow! That seems like a major story point that we need to turnaround in a couple of weeks.’” The design of the forest dwelling spirit had to be dramatically changed. “When we originally started out with him we were given a lot of instructions of how massive and majestic this creature was supposed to be. It’s interesting you spend so much time looking at something you start to lose a little of perspective with it. When we got the plates brought in and put him into the shot, we looked at it and went, ‘Oh, my God! He looks horrible. He’s massive. He looks like this giant cow.’ We had to go back to the drawing board. We pushed everything way too beefy and bulky; we had to dial back a lot of it in order to bring out some of that majesty again. When you don’t have the plates to put them in perspective at the beginning it’s easy to get steered off onto the wrong path. When you drop that character back into the world he’s suppose to be in it can be jarring.”
“There were some concept artwork and a teaser/trailer Rupert and Cedric had produced to understand the world they were going for,” states Pixomondo Visual Effects Supervisor Bryan Hirota who collaborated with colleague Geoff Scott on the project. “By the time we were talking with them they were already on location so we dealt with Cinesyncs and conference calls. We started with the material they presented to us and created tests that we would upload to them. We would then discuss what worked and what didn’t work about those tests and continued in that manner.” Dealing with two visual effects supervisors did not complicate matters. “If Cedric and Phil were out of sync creatively I could see it being an issue, but that definitely wasn’t the case here. They both seemed to have a common understand of what Rupert wanted to achieve and how best to go about it.” The compressed work period loomed over the work. “With such a tight post-production schedule the biggest thing we could do up front to try and facilitate the back end was get as much look development out of the way as soon as possible. We went back a forth quite a bit with Cedric and Phil while they were on location to identify and narrow down as much as we could the various looks we were creating. There was a teaser that was coming out before the end of the year and that was to feature some shattering knights so that was one of the main things to figure out early.”
“The most concentrated visual research we did focused on the shattering knights,” states Bryan Hirota. “We started with some concept art provided by production which gave an indication of the basic effect and the feeling that they were after. There was also this basic idea that the knights would shatter into shards. From there we moved into creating various effects based motion tests to work on exactly how these knights would shatter.” The shattering knights proved to be the biggest challenge. “We knew that the knights had to break into these shard point shards because that was a specific design feature which was desired. Also, the shatter was caused by strikes during combat where the knight was moving, so there needed to be some kind of inheritance of motion. We used thinking particles in 3ds Max to create a procedural system that would accurately break up the knight along the strike points. This allowed us to change animation of either the knight or its attacker and then run a new breaking simulation. This flexibility to tweak, adjust parameters or animation, and re-simulate the effect was the key to the results. Whenever you’re dealing with simulations it’s always a roll of a dice to some degree as to whether or not you’ll end up with results you are happy with, and our flexibility and iterative system allowed to us to run enough simulations to get the results we wanted.” A concerted effort was made to prevent the shattering from becoming repetitive. “This was a concern of ours to try and focus on some differing aspect of the shattering effect for each of the knights who get killed as to avoid repetition; we did this by staging and framing the action in a manner that was visually interesting to the shot and where we had flexibility to get the strikes to hit differing places.”
“The opening battle sequence had some additional challenges aside from the aforementioned shattering knights,” says Bryan Hirota. “There were a number of wide establishing shots that required doing a bit of terraforming to change the grounds into burned out waste lands with the requisite fire and smoke plumes. Large numbers of digital black knight soldiers were required and they needed to be able to fight with the king’s army as well. In additional to increasing the ranks of the black knights we also filled out the kings army on horseback … adding in CG knights on digital horses to effectively triple or quadruple the size of his forces.” Practical effects had to be blended with digital ones. “The augmentation of the explosions in the sand and the real fireballs thrown by the trebuchets required a mixture of creating CG simulations/renders that matched their practical counterparts and also blending in pieces of the practical elements where we could. Production had shot some really good reference of the practical fireball elements provided us a nice hard target to aim for with our cg elements. When the balls would hit the sand they were to explode. They had some on-set pyro that was buried in mounds of sand. We needed to paint out all of the sand mounds from the beach and then augment the sand explosions to put in a fiery impact component that wasn’t really present in the plate photography.”
“They would define the fantastical aspects of it up front,” remarks Bryan Hirota. “For example, ‘we have these knights that shattering into black obsidian shards when struck’; however, from there the desire was to execute the effect in as realistic manner as possible. It was accepting the one or two conceits and then working out how it would work out in the real world.” The contribution of Pixomondo went beyond the Prologue. “In the final assault on the castle we had to extend and augment the troops that have amassed on the cliff overlooking the castle. We used additional takes and replicated the troops and added in some atmospheric wheat chaff. As the troops rode towards the beach, we again employed our digital Calvary to make the size of their force larger. Whenever there was evidence of human contact on the beach we cleaned it up – including the pyro mounds discussed before. We added in the fireballs as discussed above and then again as they got closer we added in numerous arrow salvos and their subsequent interaction with the attackers. Once they arrived at the gate and entered the courtyard we provided backgrounds for the area just out side the gate and looking out of the portcullis.”
“We had a very specific brief from Cedric in the initial stages,” states Double Negative Visual Effects Supervisor John Moffatt. “We knew that we would be working on the Dark Forest and the Dark Fairies. I spoke with Cedric quite a lot in pre production mostly about the Fairies; he was very clear about what he wanted to see from them and we then set about running tests and initial designs which we presented to him on a regular basis.” Reporting to Cedric Nicholas-Troyan and Phil Brennan was a straightforward matter. “We tended to deal with one or the other most of the time up until the final delivery when both of them would have input.” When informed of Troyan’s remark about giving “the dark, cold, and ‘death’ by design” to the London-based VFX facility, he chuckles, “Cedric liked the darker type of work dneg had done on other film projects. Perhaps we are bit twisted and dark here.” Moffatt reveals, “We let the Dark Forest evolve. At the outset Tania Richards, our concept artist and vfx art director, worked up literally hundreds of paintings that illustrated ideas for populating the forest. We showed this work to Rupert and Cedric. It was a process of evolution, but nothing was concrete until we started working on the cut sequence. We continued adding and adjusting right up till the final deadline.”
“The biggest challenge was designing a toolset that enabled us realize the Dark Fairies,” states John Moffatt. “Our initial work involved sculpting a 3D maquette of the creature. We showed this as a 3D turntable to Rupert to get signoff on the design and form of the creature. Once we had reached an agreed 3d concept we then had an object that we able to have rigged so we could start animation tests. It also enabled us to have a model that we could ‘goal’ our fx work towards. The nature of the creature is that the fx flock together to form the creature, then shatter and reform when the creature has been struck or is required to move quickly. There was a period of around three months in which the team led by FX Supervisor Aleander Seaman and CG Supervisor Dan Neal worked with Pablo Giminez and Pete Kyme to develop a system and set of tools that was robust and flexible enough to hit the ground running when it came to actual shot production. Before we shot the live action for the sequence we created a sequence of three shots that we fully rendered, lit and composited that clearly illustrated what the creature would look like and the way it would move. The Dark Fairies are composed of thousands of glass pieces. We modeled a group of different shards that were signed off by the client. The shards were all given a goal of the model of the fairy that we created. The simulations were driven by the movement created by the animation, and the fx flocked the shards towards or away from the creatures depending on the action required.”
“The shape of the digital statute was fully designed but they weren’t sure about the way he was to appear in frame,” states The Mill Visual Effects Supervisor Nicolas Hernandez who was responsible for creating the Mirror Man. “When Cedric and the director Rupert Sanders made a reel for the film to sell their idea to the studio they shot a liquid black paint in slow motion coming towards you as the Mirror Man. The first brief to us was to have something liquid, fluid, and elegant not something that would morph like liquid metal; they wanted it different than Terminator 2. The first day Cedric asked us to do a practical studio shoot in a little studio across from where we work. We shot a lot of material of different liquids and motion. We shot them with a Phantom camera at 1000 frames a second to get some shapes and motion. After the studio shoot we reviewed the footage with Cedric and Phil; they wanted us to replicate some of the effects on the computer.” A lot of experimentation was done to find a way to control the liquid emerging from the mirror. “The best result was cloth simulation and the whole effect evolved from what they liked. We had to push the whole effect into a combination fluids and cloth.” Another movie by James Cameron, which features a pseudo water pod, is more representative of what The Mill wanted to achieve. “The effect in The Abyss [1989] is elegant and has a lot of emotion; it has a good interaction with the main character. Terminator 2 is more in your face with a lot of shots of technical morphing. When the Mirror Man was fully formed he had to interact with the Queen, Charlize Theron [Monster], with subtle movements; he’s talking but doesn’t have eyes or a mouth. We couldn’t have him moving too much. We had to sell the animation via subtle movements.”
“We had to do seven shots for the trailer,” remarks Nicolas Hernandez when recalling a task which occurred early on in the production. “The Mirror Man is talking over the trailer and introducing the whole character in the film; that was a bit of a struggle. Afterwards we had a lot more time to polish the final shots in the film.” 20 shots were made in total. “The biggest challenge is the first sequence where the Mirror Man comes out of the mirror. We had to time the simulation of cloth which is quite hard to control.” The cloth simulation was created in Maya using ncloth. “We took the geometry generated from the cloth simulation into Houdini to relax the cloth so to make it look a bit more liquid.” Incorporating the reflection of Charlize Theron into the Mirror Man required practical and digital techniques. “We designed a prop on set which was basically a human shape reflected. We put a Red camera inside were the Mirror Man was on-set. Charlize had an eye line plus she was talking to the prop and could act properly with a solid shape. We recorded her performance from the Mirror Man point of view. We used that footage in 2D to reflect her into our model. All of the reflections on the Mirror Man were from the set and the light sources were done in 3D. The reflection of Charlize had to be done in 2D because the studio wanted her to be beautiful in the reflection.” The studio request required some improvisation. “You cannot reflect properly a shape on a concave surface as it looks deformed.” Hernandez believes that the clear reflection fits into the story because “The Mirror Man is inside Charlize’s mind.”
“I had a mixed feeling to be honest,” admits Nicolas Hernandez when question whether it was easier dealing with a fully CG creation. “The Mirror Man was so close up that we couldn’t get away from a lot of detail. It’s not a big dynamic shot like Transformers where you get the robot in front of you and you barely see what’s going on. Charlize was acting with him and because all of the movements were so subtle that was quite tricky to make the CG photo-real all the time.” Motion-capture was never considered to be an option. “The actor who does the voice for the Mirror Man was on the set; he was behind the camera so we could see his performance. We never talked about doing motion-capture. We’ve got a really good pool of animators here. We had the voice and did a lot of phonetics in the way he was breathing and his head was moving. We added some subtle movement into his hands but that was all key-framed animation.” Hernandez notes, “Phil and Cedric were amazing clients who trusted us. We spent four to five months research and developing it with a small team but we still pushed the effect at the end doing variations to try out in order to get something that looks cool and different. On a big film like Snow White, it’s an iconic moment when she talks to the mirror. It’s not a lot of shots but the effect looks lovely.”
“Early in the production process we met with Rupert Sanders to discuss how we would approach the dwarf shots,” states Lola VFX Visual Effects Supervisor Edson Williams. “Lola Visual Effects has developed a technique that projects live footage of the hero actors’ performance onto a body double, and we felt it was a perfect fit for Snow White.” The two production visual effects supervisors brought their own strengths to the task. “Cedric and Phil both came from strong compositing backgrounds and their knowledge of how a shot should be assembled was very apparent. Phil is one of the top Flame artists in the industry, and the normal smoke and mirror tricks did not impress Phil. Cedric has worked on most of Rupert’s previous cutting edge commercial projects and he was very good at predicting what Rupert may like.” Williams adds, “It was very interesting working with two visual effects supervisors. Phil was easy going and supportive “a great start but we need to work on this, this and this”. Cedric was more abrupt but his notes were always concise and accurate. Phil and Cedric worked very well together, and when they had slightly different opinions on how a shot should proceed, they would discuss it until they came to a consensus. The entire process was ego free, and Cedric and Phil became a well-oiled visual effects supervision team.”
“The post-production schedule did not feel tight until we started to fall behind,” notes Edson Williams. “Initially we struggled to get a good facial performance projection onto the rubber masks of the body doubles, and it wasn’t until the home stretch that we really got good traction on our workload.” The facial replacement of the dwarfs required a team effort. “The production provided us with great visual research. We were giving cyber-scans of all the hero actors and the body doubles, and also cross polarized high res stills. The single most useful image we had was an 11x17 photo of all the individual dwarfs laid out. We would use this ‘master’ dwarf photo in the screening room every day.” Williams reveals, “The biggest challenge for Lola was trying to fit the hero actors’ face over a rubber mask worn by the dwarfs. The process was complex and it started as a plaster cast taken of the hero actor, converted into a painted rubber mask then stretched over the head of the dwarf. The resulting hero actor mask would get stretched into unrealistic proportions. Initially we stuck a new face on top of the rubber mask, but the facial ratios were all wrong. Unfortunately we had to deform the dwarfs head back to normal proportions before we could even begin the facial projections. Our face projections then had to include the hairline, ears, neck and face. This was the most difficult work we have ever done.”
Lola Face Projection Rig Image: Lola VFX
Assisting the Santa Monica, California-based VFX facility was the Lola Face Projection rig which is pictured in a slightly different arrangement from the Snow White and the Huntsman configuration. “The chair in the center rotates on a turntable and usually would be facing the opposite direction,” explains Edson Williams. “The light panels are normally used for outdoor billboards and we can get down to f22 for our exposure. We tend to shoot around f11 so the talent does not get a tan.” Williams states, “The facial projection process was originally developed for the twins in The Social Network [2010]. David Fincher made a great effort to have the body double’s facial shape match closely with Army Hammer [Mirror Mirror]. We only had to replace the skin from bellow the hairline to the bottom of the chin. For Snow White we had to reform the body double mask back to human proportions, then seam onto painted latex instead of human skin.”
“We got involved fairly early on,” states Baseblack Visual Effects Supervisor Steve Moncur. “We were asked to come in to take some of the burden off the Art Department in terms of the design of Duke Hammond’s castle; they presented us with a couple of early sketches of the interior courtyard and showed us some construction drawings of what they would be building on-set. We were also given some castle locations which had looked been at and a brief in terms of the period and the history. It was up to us to go away and find some additional material and present it back to them and Rupert; they treated us as an extension of the Art Department in that respect.” Working with two visual effects supervisors was manageable. “Given primarily we were dealing with Phil we weren’t getting one brief from one supervisor and a different brief from another. We had the odd occasion some contrary feedback to that but that was largely because those guys didn’t have the chance to talk it through before the notes would go out to a facility. Those kinds of things happen in the course of production sometimes especially when you’re dealing with two supervisors. On the whole they usually managed to get together form some consensus of opinion before things were going out.”
“You have to have some conviction in your own vision of what it’s going to be,” states Steve Moncur when addressing how he coped with the tight deadline. “You also have to be preemptive. According to the schedule the footage was to come in September but we didn’t get any shots turned over to us until the first week of December. So you can imagine that is a three months delay in turnover in coming to us. Officially, we had to have everything done for the middle of April.” Assisting the Art Department came in handy when bridging the shorten time gap. “The only thing that slightly made things easier for us is that we were able to get involved in the early stage with the concept design. With Duke Hammond’s castle we able to put some of the resources which would have been used to do shot production into building assets. We knew they were going to be using certain types of towers and walls. We put the 3D guys to work building 3D models which then could give us a better base to use for rendering out assets for matte paintings. The other thing that worked in our favour in a roundabout way was that we worked both on the Prologue and the Epilogue, the marriage of Ravenna to Snow White’s father and the coronation of Snow White. They both took place in a cathedral throne room; that environment originally Rupert wanted to shoot on location at Durham Cathedral in the UK but it was prohibitively expensive to hire the cathedral. We went to Durham Cathedral, and photographed and lidar scanned a lot of it. We were able to rebuild the cathedral and spend more time yet again building a better base 3D model. For us, what we lost time in shot time gave us a more time to prepare assets at a higher level.”
“The cathedral was quite a challenge because Durham Cathedral is a unique big space but in some respects, it wasn’t in keeping with the small partial set that they built and the whole Norman architectural feel that the castles and the other exterior structures had,” states Steve Moncur. “One of the biggest challenges was trying to find the right balance of architecture and scale. We went through quite a lot of iterations about how big the space should feel and how high the ceilings were going to be. We had to spend a lot of time laying things out, pulling the model apart and reconstructing it. We did have something that had some of the essential elements of Durham but had the right sense of scale and feel to what the space should be if it was a real space constructed at that historical time period. Duke Hammond’s Castle was challenge in itself in getting the space right so if felt the same from both the interior and the exterior. We had a huge helicopter shot that Rupert wanted to do. It changed in our brief from being a 2½D projected matte painting into being a full 3D model and that in itself creates challenges of getting the sense of scale right; especially on a helicopter shot where it’s easy for a castle in isolation to feel like a model. In this day and age we’re use to looking at castles which are ruined and have a modern environment around them which gives you that sense of scale, like the gift shop or the car park out front; your eye is instantly drawn to that and you suddenly understand the size and scale of the structure itself. If you take all of that away your eye looks at it and thinks, ‘How big is that wall? How thick is it?’ It was a case of perseverance, playing with the colour, depth and atmosphere, and being able to introduce human scale elements into the scene to feed that information to your eye.”
“The helicopter shot that we did for Duke Hammond’s castle ended being filmed at a cliff face side in Wales,” says Steve Moncur. “They wanted us to keep the cliff face but to sculpt a path into it so you have a sense of progression and journey up to the castle as well as some Scottish Highland mountains to go behind that.” To help with the background replacement reference plates were shot from a helicopter. “They wanted us to move mountains around and compose the shot in a certain way so we had to go in re-sculpt in 3D the Scottish Highlands. We had a very good topographical scan we took into ZBrush to push and pull things around, and retexture it to get the Scottish mountains feel but constructed to fit the shot that Rupert wanted.” Arundel Castle in West Sussex, UK served as the visual reference for Duke Hammond’s castle. “It’s one of these organic castles that have a very old keep in the middle which has been added onto over centuries. We striped away the 1700s and the stuff built in the 1800s. You have to decide what the essential elements are, photograph that and look via the wonderful tool of the Internet at other types of castles in this country. They were clear that it should be a Norman castle and have a particular style. We found that style of castle is particularly prevalent in Wales and researched to find the types of details we needed.” As for what it is involved in producing a seamless set extension, Moncur notes, “90 percent of it goes down to the research that you do, and getting a good model as well as the texture and the detail right. When looking at big expanses of castle walls you can’t just take one texture and repeat it. You have to detail small sections to create a sense of variety so your eye doesn’t does not pick up on any visual repetition.”
Baseblack was also responsible for breath enhancements which take place in the Dark Forest. “Phil and Cedric had seen the frozen breath work we had done on the last two Potter films where we had to make it feel that it was cold on the set,” states Steve Moncur. “We’ve a nice setup for doing an effect to do breath coming from the actors’ mouths; it’s driven by the soundtrack. We get the soundtrack from editorial, feed that into the macros we have written and it does a particle emission in 2D. We can apply different mist effects via this particle expression and play with the amplitude of the sound to get more punchy breath emissions; if they’re more agitated or want to emphasize a particular word or vowel we can go into the wave curves of the sound and amp them up to get different effects.” An old software program had to be resurrected. “For us, the big challenge for that was oddly enough was digging that macro out of our library and getting it working again. We had done all of the shots previously for Potter using Shake. We don’t use Shake anymore in the studio. We wanted to transfer the macro and rebuild it in Nuke; that didn’t work so we had to get Shake back out of the box again and onto the system. Most of our 3D artists are used to using Nuke; getting them to get their heads back into using Shake again was a challenge in itself. Once you’ve got it working creatively it’s about finding that right balance of using faint white breath against a snowy forest background so you can see it there. You end up having to do little roto shapes around the mouth so it doesn’t bleach out the lips. The trick with it we’ve found is to be much more subtle with than you might expect. It is easy to be heavy handed with it. It’s quite a soft and gentle touch to get it looking naturalistic.”
“When we came on board in pre-production there wasn’t yet an approved design for Snow White’s Castle,” states BlueBolt Visual Effects Supervisor/Director Angela Barson. “We needed to start building it in order to be ready for shot turnover at the end of the year, so we took on the design of the castle as well. When we were getting close to design approval we sent James Sutton, one of our modelers, up to Pinewood so he could be as close to the clients as possible.” The approach allowed for quick changes to be made, thereby, speeding up the design process. “The final sign off on the design of Snow White’s Castle happened about a month later than we’d hoped, which had a big knock on effect to the rest of our schedule. We were getting shots turned over to us in January that required the full CG castle, but the CG castle asset wasn’t completed as the design had only just been approved. The castle was such a large CG build, seen in so many different lighting conditions and states of disintegration, that it wasn’t possible to make many shortcuts.” The other major job assigned to BlueBolt was more straightforward. “In contrast, the design of the Royal Village was known early in the process, so we started the CG build of the royal village while the shoot was underway. By the time we were given shots to work on, we almost had a fully built and textured CG village. On the whole, the village shots went through very smoothly.”
“The biggest challenge was creating the CG castle in the short timescale, and making it sure it could be rendered in time,” remarks Angela Barson. “The castle is such a large build and is seen from all possible angles and distances that the level of detail in the modeling and texturing had to be very high; in some ways that became the area of difficulty – getting the level of detail high enough for the close ups, but not too high for the wide shots. We textured the village in Photoshop, but when we moved onto the castle texturing it quickly became apparent that the number and size of textures required was too big for a Photoshop pipeline. We made the decision to move across to using Mari as it was far better at dealing with texturing on this large scale.” Barson continues, “The fully CG castle shots were definitely the hardest to achieve. Where there’s something in the shot to match to for look and lighting, it’s always easier to get it right and have everyone in agreement. When it’s all CG it can be open to everyone’s individual interpretation. During the battle sequence on the beach we had many shots where we were adding the fully CG castle. As always happens, the lighting conditions changed quite dramatically between some shots. The backlit, atmospheric plates were much easier to work with than the front lit, clear plates, and there’s always the challenge of trying to introduce a level of consistency between the CG lighting regardless of the variation in the plates. We added in a lot of additional atmospherics to the CG to stop it looking too clean and therefore CG. Small pockets of smoke, flambeaus, birds, sea mist and people were added to bring the castle to life and give it a sense of scale.”
“We were responsible for all the set extensions of the royal village in both the Magnus and Ravenna reign,” states Angela Barson. “This involved topping up the central houses as well as adding fully CG houses to extend the area. We also did set extensions to Snow White’s Castle courtyard that was built as a set piece at Pinewood Studios. The plates for the main wide castle shots were filmed out in Wales where no practical build existed, so we built a full CG version of Snow White’s Castle in both the Magnus and Ravenna reigns. The number and complexity of the fully CG castle shots was far higher than we had initially imagined, but it was a great challenge and we have some fantastic shots as a result.” As for the art of making invisible set extensions, Barson remarks, “We started by getting a lidar scan of the set so that we could match our CG extension exactly. We also did an extensive texture shoot of the set to use as a basis for our CG textures. This had to be done twice for each set build as they were dressed differently for the Magnus and Ravenna reigns. We built the nice Magnus reign first and then added all the destruction for Ravenna as a second pass. Creating the run-down Ravenna version involved destroying walls and roofs, adding additional staining to walls, having vines growing around the castle, blackening any timber and generally making it look more charred and neglected. This type of CG environment work is always fun to do.” The digitally crafted settings were grounded in reality. “The key was keeping the design and the scale of the detail realistic. Even though the castle is enormous, it is designed in a way that is structurally sound with a sensible scale to the stone blocks. The scale of features like doors, windows and buttresses are correct, and there is a lot of smaller detailing like vents in the roofs and timber struts to break up large surfaces. We added lots of subtle variation to the stonework, as it could never have been built all from the same stone type in a short space of time. Staining was added to make it look as though it had been standing for many years and to give it an extra sense of realism. It’s the addition of all this subtle detail that helps to keep it looking real.”
“We worked with Rupert before on quite a few commercials,” states Effects Supervisor Lindsay MacGowan who works for Legacy Effects. “He made a small five minute pitch movie for Universal which we designed some stuff for; there were some tree puppets which we built, and some elements for a melting apple.” Teaming up with Rupert Sanders for a movie was not a vastly different experience from previous collaborations with him. “He is use to doing grand scale action oriented commercials.” MacGowan and his team dealt with a small group of people. “It was Rupert, his producers, and Colleen Atwood [Chicago], who had worked on a lot of commercials with him, did the costumes; it felt like a group of friends coming together to get this project done.” The initial job for the California-company was to assist the digital trickery which would turn the full-sized actors into dwarfs; however, the major assignment was producing the armor for the Dark Army. “For the armor, Colleen came to us with a Photoshop design.” A digital sculpture was crafted under the close supervision of Atwood and Sanders which was sent off to assist the various visual effects companies.
“A stunt performer came here to the studio and we did a couple of things with him to make sure that his movements were okay,” remarks Lindsay MacGowan. “We can alter our master pieces to accommodate any type of action that he might need to do. If there is any rubbing or discomfort in certain areas we would try to do our best to alleviate that by changing the design within the costume so that they can work comfortably but still give the director the look he’s going for.” When came to achieving the proper balance between a gritty realism and the fantasy element, MacGowan says, “We showed Rupert what the process was and the type of materials we were using.” 150 suits plus some backup pieces had to be made within a three month deadline and shipped off to Pinewood Studios. “Originally with the prototypes there were a 150 pieces for each suit and eventually we managed to get that down to 75 pieces.” MacGowan notes, “It was an immense amount of work in a short period of time; it was great working with Rupert and Colleen.”
“Nvizage was given a very clear brief of the production’s requirements with regards to previs,” states Nvizage owner and founder Martin Chamney speaking on the behalf of the London-based VFX company. “Using a virtual camera system operated by Cedric, DP Greig Fraser [Let Me In] and the Director Rupert Sanders, meant the opportunities as well as constraints of sets, locations and even CG characters could be experienced first hand. Cedric had a clear vision as to what he wanted and his involvement in recording mocap and shooting sequences meant his ideas could be envisaged more easily.” The work had to be completed sooner than expected. “Nvizage was originally instructed to complete previs for the Troll Sequence and various battle scenes which would take approximately six weeks. There was a possibility that the Troll sequence might be omitted, so we placed this previs sequence on hold. Finally, it was confirmed that the scene would be included in the film two weeks before the end of our schedule. Getting the digital assets prepared, all the animation completed, the entire sequence shot using the VCS and all rendering completed before shooting started took a lot of planning and a super efficient pipeline.” Various types of visual research were conducted. “We created previs assets based on everything from stylized maps, concept art, location photos, Google Earth data and Art Department plans. It was often up to us to figure out how all these elements pieced together, especially CG sets which were based on real locations, such as Ravenna’s castle and the village.”
“The main sequences Nvizage prevised were the battle scenes in the forest, the final battle on the beach and the Troll Bridge Sequence,” states Martin Chamney. “The main challenge we had for the battle sequences was ensuring we could include all the animated horses and riders in motion builder whilst ensuring we could play the animation in real time while running the VCS. With the Troll previs one of the main challenges was taking existing mocap data and stitching it back onto the characters so they were one seamless animation.” The biggest challenge was completing the Troll Sequence within the shorten time frame. “We decided to take what was essentially one extremely long and complex sequence to be shot using the virtual camera, and break it into four parts. The work was divided between various artists making the entire workload more manageable and efficient. This was especially useful when animation changes were required making it easy to adjust one part rather than altering and offsetting animation for the entire sequence. We could shoot one section of pre-vis whilst animation was being refined for others, working in a non linear manner not requiring the start of the sequence to be completed before we began on the end.”
"The key to producing effective previs is building a good relationship with the director and DP, and other HOD’s and understanding their mindset and vision for the film," observes Chamney. "Involving these people in the creative process means that they are more likely to plan the shots that they want. The previs then provides them with the information and confidence to achieve the same shots in real world. Working this way avoids wastage and we strive to produce effective, realistic, and achievable previs that always gets used and seen in the final film. Much of our skill is less about making glossy over polished content, but trying to focus on what previs is really for – storytelling, timings, layout, composition and great camera work.”
“Cedric, like Rupert has a great eye for detail,” states The Mill Creative Director Henry Hobson. “Even though he was deep in thousands of other visual effects shots he would always pop in to chat about the titles, bringing some great ideas and thoughts into the process.” The Los Angeles facility for the UK based VFX company was responsible for creating the Opening and End Title Sequences. “We decided to take all the costumes from the film and bring them to life in one final battle between good and evil. Colleen Atwood's costumes are as staggering in person as they are on film; she does an amazing work in creating details that appear like real steel, bronze and chain mail even to the most discerning eye. We photographed the costumes in extreme details and in the midst of action with some raking light; this formed the bed of our design process. With a lot of treatment and processing we began to explore typographic options. Building rough edits with still frames before we got the green light to shoot the final piece with the phantom. The final processes were amazing combinations of our core team Andrew Proctor, Eugene Guaran, Yorie Kumalasari and Ed Quirk. Manija Emran joined the team to help steer the typography and creating a completely custom typeface.”
“We wanted to be able to blend between the two worlds, CG and in camera, without the viewer seeing the joins, and the Phantom camera allowed us to do this,” says Henry Hobson who had to incorporate bird footage with the digital effects. “The raven was easier to work with than some of the stand-ins! Shug was almost like a kitten; it was so playful, fun and friendly. On any command it would flap, fly and jump. The shoot we had with him was very short so it was great to have a creature so well trained.” The shattering effect served as a key design element. “Rupert initially requested a flint like texture, with ripples of organic shapes in the broken pieces but as we got closer to the finish line, we eventually had to even up the shapes and make them more graphic, as the realism of the ripples took the viewer away from the hard cut graphic nature of the sequence. We only had a couple of weeks to do this in, so it helped to work alongside Pixomondo and add our own level of extra detail that was needed for our macro shots.” As for the creation of the custom typeface, Hobson remarks, “After trying several fonts we had trouble finding that perfect feminine font with a hard edge. When Manija Emran came on board she expanded upon traditional type selection by introducing the possibility of doing a custom font. Manija took a font not traditionally used on screen because of its fine line work, and began to customize it. We ended up with a new font with several options per letterform, which allowed for each name to have its own personality, without being too stylized.”
A current Hollywood trend was purposely avoided when making Snow White and the Huntsman. “It was already a tough schedule if there had been any 3D involved in this it would have made it close to impossible given the very short time frame,” states Phil Brennan. “All three of us are happy with how most of the movie turned out. The opening battle, those shattering effects look great, and the dwarf stuff worked out really well; these were things everyone was skeptical about how all of these techniques were going to work and they all came together in the end.” Pixomondo Visual Effects Supervisor Bryan Hirota remarks, “This show was quite challenging due to the number of shots [250+ish] and the short post production schedule but overall it was quite an enjoyable show. The creative team of Rupert/Cedric/Phil all presented us with such a unique and clear vision that it really was a pleasure to help them achieve it.” A lot of time was spent creating nuances for the movie which will never be noticed by an audience. “I’ve certainly been to films where I’m watching my work go by,” says Rhythm & Hues Visual Effects Supervisor Todd Shifflett, “in a blink of an eye there goes something you know that you’ve spent weeks trying to fix and you want to be able stop and tell the theatre, ‘Wait! You don’t understand how difficult it was!’” Asked about his favourite scene, Cedric Nicolas-Troyan replies, “It’s hard because all of the big VFX sequences have such a different vibe and they are all my babies in a way so I am proud of them all.”
Technical support provided by Csaba "Hungarian Wizard" Szabo and Gary Collinson.
Many thanks to Cedric Nicolas-Troyan, Phil Brennan, Todd Shifflett, Bryan Hirota, John Moffatt, Nicolas Hernandez, Edson Williams, Steve Moncur, Angela Barson, Lindsay MacGowan, Martin Chamney and Henry Hobson for taking the time to be interviewed.
“We wanted to keep the fairy tale aspect somewhat grounded,’ states Cedric Nicolas-Troyan. “The first thing was not trying to explain everything and anything. The second thing was to avoid any type of light, glowy or over-the-top effectsy FX as much as possible. We tried to root any type of fantasy aspect so it is not realistic but still feels real in the world we created. We also tried to keep the design unified and away from a patchy all over the place type of feeling, especially, when you have to cover so many different types of visual effects.” Phil Brennan [Terminator Salvation] notes, “Even though there is the underlining fairy tale theme to it, it’s a gritty adventure action film. A lot of that comes from Rupert. Rupert didn’t want things to look over-the-top; he wanted things to have a gritty dark believability about them so we were always leaning in that direction.” Nicolas-Troyan observes, “It's funny how everybody imagines Snow White as the Disney character. The original tale is much more vague and dark.” Brennan sees more of a connection with the fantasy trilogy helmed by Peter Jackson than the medieval story directed by Ridley Scott. “The Lord of the Rings [2001 to 2003] comes in because there’s a lot of scale. There’s a lot of journeying so it has that same epic feel to it. Robin Hood [2010] I’m not so sure. We’re drawing on anything that has tons of battles.” As for what will make this version of Snow White distinct, he remarks, “That comes from Rupert himself; he has a unique vision and visual style so bringing that into play, at least visually, gives the movie a unique look.”
“A Western European Raphaelite kid was the ideal image for Rupert in regards to what the Good Faeries looked like,” explains Todd Shifflett. “He had chosen two children to play the part of the faeries and generate facial motion-capture for us. We modeled our faeries after those two children and tried to use that facial motion-capture wherever we could. Unfortunately at the time there was no scripted section for the faeries in the film so they were getting a random action that doesn’t work so well when you want to go back later on and put something specific in. You need to work in a certain direction if you’re going to use motion-capture.” To compensate digital keyframe animation was utilized. “We used that motion-capture as a reference for our animators to be able to go in and see the real details of the organic nature of the animation. These tiny details might be a flinch of the eye or a quick firing muscle that doesn’t seem like to do anything in particular; all of these little things add up to feeling of a naturalistic movement, whereas if we just have a lot of nice blend shades it might provide nice animation from an animated movie situation but it wouldn’t necessarily provide an organic natural animation. What we wound up doing is driving our animation by hand and going back referring to sections and go, ‘Look at what’s happening here. We need to get this feeling of the cheek movement happening in our own hand driven animation.
“We worked with Rupert before on quite a few commercials,” states Effects Supervisor Lindsay MacGowan who works for Legacy Effects. “He made a small five minute pitch movie for Universal which we designed some stuff for; there were some tree puppets which we built, and some elements for a melting apple.” Teaming up with Rupert Sanders for a movie was not a vastly different experience from previous collaborations with him. “He is use to doing grand scale action oriented commercials.” MacGowan and his team dealt with a small group of people. “It was Rupert, his producers, and Colleen Atwood [Chicago], who had worked on a lot of commercials with him, did the costumes; it felt like a group of friends coming together to get this project done.” The initial job for the California-company was to assist the digital trickery which would turn the full-sized actors into dwarfs; however, the major assignment was producing the armor for the Dark Army. “For the armor, Colleen came to us with a Photoshop design.” A digital sculpture was crafted under the close supervision of Atwood and Sanders which was sent off to assist the various visual effects companies.
0 comments:
Post a Comment