Go To Black: The Making of Iron Man 3

“What Iron Man 3 gave us was the opportunity to play around with a brand new suit,” remarks Christopher Townsend.  “This is now the Mark 42.  The last we saw was in The Avengers which was the Mark 7.  Tony has moved on with his suit technology and that afforded us the luxury of being able to play around with the way the suit moves.  It was a much more delicate suit in terms of the tiling and the entire structure.  It’s what we call a prehensile suit meaning that it was able to form as different pieces, attach to the body as individual components and was remote controlled.  The Mark 42 was a fun challenge.”  An army of armoured suits had to be constructed.   “We built a total of 41 different Iron Man suits of one sort or another.  We did the Iron Patriot and the older suits, and then a whole slew of new suits people haven’t seen before; that’s what we see in the end battle that Weta worked on which was this vast Armageddon type of experience.  The fun part was to design each individual suit.   Whether it was the weaponry, flew apart, how it looked and what its purpose was; whether it was a space suit or an underwater suit.  Even though each of these suits only have small cameos and on-screen for maybe two or three shots each one had to be designed and built at a high quality so that when you saw it you believed it to be part of the environment and a character in the film. There was a tremendous amount of effort and work to make sure that they hold up on-screen.”  Suit concepts were merged together so to create more variety.   “We built about a dozen hero suits which had unique body parts and then from those 12 suits we built another 20 odd suits that were a mix and match of those pieces.  The arms would be the same from one to another but the chest or neck piece or head or legs [would be different].”

“We have a sequence where Tony Stark’s [Robert Downing Jr.] house is blasted off a hill,” recalls Christopher Townsend.  “When Dan and I talked about it we talked about doing miniatures and shooting practical plates then we found out we couldn’t shoot the practical plates at the real location.”  The demands of the screenplay required that the Malibu, California setting be digitally altered.  “Because of the action of the house being blown off the hill we needed a sheerer cliff.   We would have to re-sculpt the cliff, make the water deeper because we needed the story point of the house going deep down into the ocean but in reality it is a rocky bay which is quite shallow.  Going much deeper we would have to re-sculpt the entire surface and make it looked like it had re-emerged.   We would have to have helicopters which could fly with these huge gun armaments with door openings which was practically something we couldn’t shoot.  We also wanted the shot at a particular time of day, what was magic hour, not quite so pretty and ominous.  There were so many constraints with what we would have to do that we looked at it and said, ‘It doesn’t make sense.  We’re going to have to have some digital shots so let’s bite the bullet and make it all digital.’ The House Attack Sequence involved a lot of advance planning.  “We turned the shots for that sequence over to Scanline before we even started shooting the principle photography because we knew they would be all CG.  They worked on some of those shots for well over a year. It’s exciting stuff as they pushed the envelope on figuring out how to make something that big and dramatic look real and maintain that sensibility at same time.  Dan agreed that when we were shooting inside the house and had the actors running around we needed to shoot something as a practical base so he created along with Bill Brzeski [The Hangover], the production designer and his art department, a full-size living room section of the house with draughtsmen; they put the whole thing on a gimbal and at a chosen moment the whole house split in half physically with our actors [Robert Downing Jr., Gwyneth Paltrow and Rebecca Hall] acting in the environment.  Having that as a practical base was the key to giving it that sensibility of being in that environment.”  In order to produce the militarized news helicopters, the circumstances surrounding the antagonist of the movie needed to be taken into consideration.  “The whole thing is a process as you try to figure out, ‘How would they physically build them and because of their funding what could they afford to do?’”


“We ultimately had 17 different companies working on the show including our small in-house group of artists,” states Christopher Townsend who was responsible for 2018 visual effects shots.  “It was a lot of shots and a tight schedule.  We originally had a reasonable post-production schedule.  They had pushed off shooting the movie and then right in the middle Robert Downing Jr. [Wonder Boys] injured his ankle doing a stunt.  We went on hiatus for six weeks.  When we finally wrapped photography we had 19 or 20 weeks left before we had to be finished.  That’s a tremendous amount of work to do in a short amount of time so we moved quickly and needed to spread the work.  Unfortunately, a couple of companies we were working with went into bankruptcy.  With the limited amount of time we had, additional visual effects companies had to be recruited for the project in particular for the End Battle Sequence.  We needed a powerhouse company that could take on the volume of work and be able to work on it with a huge talent pool to get the work done.  I wanted to pick a company that had done this type of thing before.”  Weta Digital proved to be the answer.  “They did over 500 shots in a period of about three months and it was a full digital environment, full medium and close-ups of digital characters, explosions and pyro, effects type work.  There were 40 odd suits they had to build [as well as the Extremis characters, people who glow from the inside].”

“The delay in shooting caused by Robert Downey Jr.'s injury had the knock on effect of shortening an already short post schedule,” states Weta Digital Visual Effects Supervisor Guy Williams.  “We came on to help distribute the work load out so that the show could get done in time without any company having to drop quality.  We had the advantage of having recently finished Avengers and I was able to get the same core team onto Iron Man 3 [with some excellent new additions].  We set about doing the work by being very realistic of what would need to be done [no hopeful expectations, just cold hard reality].  We used this to make a plan that was aggressive but would also cover all possibilities.  This way, we could make sure we could do what Marvel needed without having to rebalance our plans.  We future proofed our plan in essence.”  Understanding the needs of the production was not an issue.  “Chris was excellent to work for.  Between visits to L.A., the set, and regular conference calls, we were able to make sure that we were all aiming at the same vision.”  Williams was a member of the Oscar-nominated visual effects team responsible for The Avengers. “On Avengers, there was one Iron Man suit.  On this show, there were 36.  The complexity of the work we had to do on this project was an order of magnitude higher then the work we did on Avengers.  There were more digital doubles, more all CG shots, more characters and larger effects.  Avengers was a great show to work on and went a real long way in preparing us for Iron Man 3.”


“We started with excellently fleshed out designs from the Marvel Art Department,” explains Guy Williams.  “We then could build upon those designs as needed to make the suits work better for animation.  The creative team at Marvel knew that we understood their designs and would be faithful to them as we modified them.”  Real military conflicts helped with the fictional one.  “For the final fight, we referenced a lot of war footage.  The desire was to make it always feel like there was a battle raging all around the viewer but not have the battle takeover the focus of the shot [unless the battle WAS the focus of the shot].  This is a lot trickier than it sounds.  It is way easy to set up an awesome background action that is so awesome, it draws the viewer into the background and away from the storytelling.  We had to do things like constantly break the edge of frame and offset the timing of the action to not sit well in the cut.”  Having to deal with an army consisting of individually distinct armour suits was helped by the type of shots required for the sequence.   “We worked to make sure that the colours and the shapes allowed us a lot of variation visually at multiple distances.  But the truth of the matter is that we only had to sell it a bit as many of the shots dealt with the suits in distant silhouette.”


“We referenced a lot of jet and helicopter videos to work on the sense of speed and weight to the suits,” remarks Guy Williams.  “We wanted to make sure as the suits flew, they banked properly, which is to say they slipped in the turns.  It was important to make sure that the suits didn't feel like they were on rails.  We also tried to deal with the exposure levels with an eye towards high energy events shot at night.  We let some of the effects blow the frame out.  In general, you try and figure out what a good analogue of the effect is that is real and you use some of the visual traits of the real thing to build the effect on.”  There were no creative concerns in depicting Mark 42 suit.  “Not really.  For us, it showed up just long enough to get blown up.”  Practical and digital effects needed to be combined seamlessly.  “We had to do a large number of set extensions from green screen sets to our digital location build.  That is about the extent of integration.  Most of the effects work we did was digital.  We use a lot of fluid simulations in our effects work.  It allows us to better control the look and the timing of what we are after.”  Sharing assets with the other visual effects vendors assisted the development process.  Digital Domain did some excellent work on the models they gave us.  We had little in the way of technical clean-up.  The issue for the most part was a scheduling one.  We had to make sure that as we shared assets around, we were doing it in a timely manner so as not to delay anyone else's schedules too much.”


“At the end, we were given the challenge of creating the Lava God version of Killian,” states Guy Williams.  “This is where Killian comes out of the fire at the end and has one last face off with Tony.  Guy Pearce [L.A. Confidential] had grown a beard so we would need to replace the lower part of his face.  We decided early on to bite the bullet and go full CG on the shots to get the best result.  We had to get a talking character up to spec and final all of the shots with heavy effects in less than three months.  That was a fun challenge.  We plussed up our digital double of Killian and did an extra FACs session during the reshoot to make sure that our digital face would match faithfully to Guy's performance.  Once we got to the point we were happy with the facial and the rendering of Killian, we started layering on the effects passes to make him look like he was heavily damaged and trying to regenerate.”  As for whether there is a particular shot or sequence he is proud of Williams remarks, “Particular?  Nope.  I am proud of all of the work the team here did.”

  

Digital Domain was originally the cornerstone vendor.  “I was impressed by their work on Real Steel [2011] and loved the reality of what they brought to that show," states Christopher Townsend.  “It started off well but unfortunately there were separate issues that changed things so we reduced the number of shots we did there.  What we did do there was the Barrel of Monkeys Sequence, several other sequences in the snow and we had an Extremis character at the Hollywood Chinese Theater.” Digital Domain Visual Effects Supervisor Erik Nash notes, “One of the things unique about Iron Man 3 is the number of suits,” states Digital Domain Visual Effects Supervisor Erik Nash.   “One of our tasks from the outset was to build those suits digitally and that task varied quite a bit suit to suit.  One of the suits we had to build was the centrepiece of the movie, what is referred to as the Mark 42 suit and that one was pretty well fleshed out by Marvel’s visual development team in terms of illustrations.  They had also done a fair amount of modelling.  Now the modelling that they do is helpful in terms of giving us something to look at from any angle.  We basically have to redo it from scratch to have that model work in our pipeline and rigging paradigm but that one was pretty well nailed down in terms of what it looked like.  Similarly there was a practical version of the suit that Legacy Effects built which we could photograph in terms of what the material qualities were.  That was one end of the spectrum.  The far other end of the spectrum is some of the many suits which appear in the Third Act only existed from Marvel as illustrations.  Usually a front and back view but in only a few cases there was only a front view.  We had to extrapolate and figure out what the back of that particular suit should look like. But in all of those cases we’re starting without a 3D model.”   Helping to make the suits unique was referencing them to familiar objects.  “There is a suit designed for outer space and it uses a space shuttle colour scheme, the undersea suit is somewhat reminiscent of one of those deep sea suits, copper material as well as blue green tarnished finish, and the stealth suit had the charcoal grey look of the stealth aircraft that the Air Force has.”

Erik Nash brought the added experience of jumping over 1,300 times from an airplane to the dramatic Barrel of Monkeys Sequence where Iron Man has to rescue passengers falling out of seriously damaged Air Force One.  “That was one of those sequences that Chris and I talked about early on, reveals Nash.  “Coincidentally Chris didn’t know it at the time that I back in my younger days was an avid skydiver.”

The first hand experienced came in handy.  “It was great to be able to understand the visceral feel of being able to skydive and also the physics,” notes Christopher Townsend.  “With visual effects you need to understand the physics and science of things often in order to make it look real.  It’s one of the things at the beginning you look at the boards and animatics, and go, “How do you do this?’  Obviously going into a full green screen mode is the first place where your mind goes and people on wires.  We’ve all seen movies where that has been done.  It was evident to all of us that we didn’t want to go that way.  We wanted to make it more real than that.  There were several things that came to me about what makes something look real.  A lot of it was the camera work.  We’re all used to YouTube and online where so many people post their own videos of themselves skydiving.  One of the things I noticed is that all of the lenses are the same because they’re all shot with a headcam or iPhones.  As soon as you start changing the lenses going for a wide angle to a long lens it takes away that visceral reality.  You suddenly say, ‘Oh, that’s being filmed.’  We tried to shoot everything on an 18mm lens.  Another big thing was when you have people shoot someone falling the cameraman is falling with the person.  It’s not really interesting.  The only time it gets interesting is when you see people floating away from you or towards you.  We did a lot more of that stuff which made it a much more exciting sequence.”

“Chris from the outset was determined to shoot as much as possible as practical free fall photography which I backed him up a 100 per cent,” states Erik Nash who makes a cameo appearance as a nearly overturned boater in the high flying rescue operation which concludes with Iron Man dropping his human chain into the water.  “We knew that we would have to do a lot of sky and land replacement given the parameters of the sequence in that when they fall out of the airplane.  They’re up at 30,000 feet and they go all the way down to sea level.   All of the actual free fall photography had to be achieved in a fairly narrow altitude range of 12,000 and 3,000 feet and the fact that we shot on the coast of North Carolina but it had to look like Miami.  It was compounded with the variability of the weather and the time of day.   They shot over a span of eight days starting at seven in the morning and going to about four or five in the afternoon.  We had a huge range of weather.   Everything from completely overcast to cloudless skies, the sun low on the horizon or straight overhead and it is all intercut through the sequence.  They did a day of free fall photography here in Southern California, seeing what ultimately it was going to take to replace these backgrounds.  We planned on doing a fair amount of that.  We didn’t plan on doing as much as we did.  I can count on one hand the number the number of shots in the sequence where we didn’t replace the entire background.  It was an incredible amount of painstaking rotoscope work and tedious edge work because these people who fall out of Air Force One are wearing civilian and military wardrobe which is flapping in the breeze, their hair is all over the place and that free fall velocity, the amount of blur from frame to frame is substantial.  Finding the edges of these people to be able change either the sky or the land behind them was an unbelievable amount of man hours.”

“We created a virtual Miami that we could see from as high as 30,000 feet down to about 50 feet in nested resolutions,” explains Erik Nash.   “As we got closer the resolution in this virtual environment got higher and higher so that over the course of two minutes of screen time which is accurate in terms of how long it would take to fall from that altitude.  One of the first things we did was once they turned over the cut sequence we plotted the shot in a time versus altitude so that we had a constancy and logical progression to where we were altitude wise over the course of the sequence.  We varied a little bit from the reality of it in that when you’re falling at a 100 or 120 miles per hour at 30,000 feet there’s virtually no sense of getting closer to the ground at all.   At the beginning of the sequence we’re falling faster and as we get low to the ground where your velocity is ridiculously fast we slowed it down.  Our range of free fall speed probably ranged from about 200 miles per hour at the high altitude to down about 75 miles per hour as we got close to the ground.”  To make the moment believable a high frequency camera shake was added to the imagery.  “The buffering of the 120 miles per hour airflow is something that anybody who has seen any free fall photography is familiar with and if you don’t have that all of a sudden it looks fake even if everything else is done perfectly.  In a couple of all CG shots where it didn’t start out from a photographed plate we mimic that high frequency camera shake so it has that same flavour that all of the real free fall photography had.”  Digital replacements served to resolve continuity issues.    “By replacing the backgrounds completely we were able to get a continuity progression that is completely accurate and true to what is happening.”

Recreating the aircraft that transports the American President was more complicated than expected.  “We’ve done lots of jetliners over the years, so we thought, ‘It’s just another 747 with a different paint job,’” says Erik Nash.  “But in looking at all of the reference of Air Force One it quickly becomes apparent that it is by far the cleanest airplane around.  It seems that when it’s not fly they’re cleaning it.  When we built and textured our version of Air Force One and put it into some plates it looked fake.” Some creative license was needed.  “We wound up making our Air Force One dirtier than the real thing.  Adding oil streaks and panel seam dirt to help make it look real.”  An action sequence takes place inside the presidential vehicle involving a fight between Iron Man and the Iron Patriot which involved a disappearing and reappearing element.   There was a Santa hat that was in.   It was out.   It was in and they shot it with James Badge Dale [The Grey] who plays Savin in the practical Legacy suit with the Santa hat.   Ultimately they said, ‘Santa hat is out.’  We needed to get rid of the Santa hat.  ‘Okay, we could replace his head.’  It was easier for us to replace Iron Patriot completely rather than just replace his head.  We did that on a fair number of shots where the plate had the performer/stuntman in the suit interacting with the other actors.”  The real footage was invaluable.   “We would always have a practical suit brought in, and shot in that environment and in lighting condition as reference because that is by far the most useful and accurate lighting reference you could hope for as you now have something you can put side by side.”

“There’s a scene after Tony’s house gets attacked and falls into the Pacific Ocean, he flies across the country and crash lands,” states Erik Nash.  “In the story it’s rural Tennessee although we were shooting in North Carolina.  The whole movie takes place around Christmas while we were shooting in August.  Shooting in the woods 80 to 90 degrees, 95 per cent humidity, August in North Carolina but making it look like December, snow on the ground. Dan Sudick and his physical effects team created a patch of real snow for us which took days because it melted so fast.” Outside of the area created by the special effects team everything else was a virtual environment including the high establishing shots prior to the crash landing, thousands of evergreen trees covered with snow, the pickup truck that Tony Stark has a near miss with and a deer on the road.  “The snow was not that big a deal.  We had to add breath.  Luckily we had done that in the Arctic scene in Pirates 3 [2007] so we had experience with that which worked great.  We had done a ton of snow in Grinch [2000] but it is always tricky because it takes place at night.  You’re making it hopefully believable in moonlight with falling snow and it’s all subtle stuff that is necessary to make it look real.  We couldn’t actually shoot the wide aerial shot where he is coming in to crash land at night because we would have to light up literally hundreds of acres of wilderness.   It was an interesting lighting challenge but in the end we’re happy on how it turned out.  It’s one of those things which are not obviously visual effects.”

“We went to Scanline in California to do the House Attack Sequence,” remarks Christopher Townsend.  “They’re one of the best companies in the world doing massive simulation work and water, particularly both above water and underwater, and I’ve admired their work for a long time.  The House Attack Sequence required a tremendous amount of effort to bring things up to a level of photorealism and to see water in a way that we didn’t feel we could shoot elements realistically.  We couldn’t shoot the elements big enough to make them work in terms of our time and schedule so we needed to rely on CG; they step up to the plate and did a tremendous job.”  Visual Effects Supervisor Bryan Grill who works for Scanline VFX recalls, “Victoria Alonso at Marvel was instrumental in getting Visual Effects Supervisor Chris Townsend and Visual Effects Producer Mark Soper [The Fountain] to come to Scanline for a tour and screening of our company reel. They had a sequence already fleshed out and thought that we would be a great fit for the Mansion Attack. They were very specific on how they wanted the sequence to look considering it was going to be mostly all cg and had to look photo realistic.  I, along with co-supervisors Darren Poe and Stephan Trojanksy, felt Scanline was very much up to the challenge.”

The work on Avengers and Iron Man 3was very different,” observes Bryan Grill.  “On The Avengers we had one very large asset, the Helicarrier, which had to interact with water. And we had to fully integrate an airport runway set into the surface of the Helicarrier.  Heavy water simulations and lots of modelling details into the Helicarrier helped establish its size, weight and mass. On Iron Man 3 we had many more environments and assets. First was the Point Dume environment and the Stark mansion that sits atop it. The surrounding mansion grounds including the interior living room and garage. We also had the cliff and the underwater environment, where Iron Man along with his house sinks to the ocean floor. The driveway, living room and garage each had sets built in which we needed to augment and expand with destruction and explosions. And let's not forget Iron Man himself. Iron Man needed to be built and animated to incorporate stunt work and the suit connectability throughout the sequence.”  CG experiments were conducted. “Chris early on wanted proof of concepts that we could hit the look that he was looking for. First we did an Iron Man test in which we rendered and animated. He was able to breakdown what he liked and didn't like, which gave us a foundation for future shots. We also put together a full CG Point Dume environment, which we viewed side by side with stills photography. That let Chris know we could hit the realism which was needed for the movie.”

“The key to creating photo real visual effects always starts with real reference,” explains Bryan Grill.  “Even though there was no mansion on top of Point Dume we had Point Dume and other building and texture reference to help us. It comes down to the details. All those little things from dirt to imperfections that are not completely noticeable but you feel in the rendering. Than the lighting and how materials react or don't react to light. Is it to shiny, specular or diffuse? Making sure you have reflections, shadows, self-shadowing and are they too dark or too light or too long? All of these and many more decisions need to be made even when the render is able to give you all the real world qualities there still is the subjective view of what looks real or not.”  A year was spent in development the scene.  We received previs and animatics before the shoot began. There were shots that had been picked by Marvel that they wanted for their Comic-Con presentation. So those shots became our tent pole shots and pushed our development. It became very clear that this was no easy task. Even getting to start on the project early on we knew that what was needed to make these shots work was going to take all the time we had. Ultimately, Chris and Victoria Alonso at Marvel paid us the highest compliment, when they remarked that, ‘We hit the mark 100 per cent,’ it made everyone at Scanline feel especially proud.”

“We began with reference Chris and the production had gathered and then we started adding to that until we had enough for us to feel we could hit the mark,” states Bryan Grill.  “With the cliff, deeper water and surface changes to the cliff, we started with some concept paintings which enabled us to get answers from production before we started to model and texture. For the magic hour look Chris had taken a picture that represented the mood and time of day for the Attack Sequence. We than took that one picture and extrapolated it and came up with a completely new sky. We started with Pan and Tiles taken at Point Dume at different times of day. Using that as our base we than incorporated other details until everyone was happy with the sky. This process took a few iterations but I think we found an environment and time of day that was believable.  For the mansion and other cg environment components we had gone out and done some lighting reference pictures down on the beach in Marina Del Rey. Basically doing time lapse photography as the sun was setting to see how the light and shadow played on different building materials. As well we played with exposures so we had really good idea how if a DP were to actually shoot our environment what decisions would be made when looking into the sun and how the things looked under or overexposed. We than picked some hero images and got a buy off from Chris which really helped nail down our look.”

“The above water surface and large-scale splashes and the underwater environment were especially crucial in telling the story of Tony's house being destroyed,” remarks Grill.  “Our internal simulation software, Flowline, was used extensively for the destruction components, to create the smoke and fire, as well as the ocean water. We created extensive RBD simulations of fracturing concrete and various house materials being destroyed.  We see splashes on the water surface from cars and debris hitting the water and also see the continuation of those splashes from an underwater view as well. We started with underwater photography at different depths which showed us how much the light affected the water as we sunk deeper and deeper. We chose a depth of 200 feet below the cliff for our mansions final resting spot. The colour palette was also very crucial it needed to feel like Southern California and not the turquoise of the Caribbean. The fall off of visibility was another challenge making sure we could see what was happening at all times by using light and shadow to feature the objects we still needed to see. There were plenty of photographs taken while on set of all the set pieces, this along with LiDAR scans gave us enough information to completely reconstruct the sets to help when needed.  It's important to note that although production used an impressive breakaway tilting set, we did extensive interior work digitally; that involved set extensions rigged for destruction, as in the case of the ceiling almost crushing Tony as well as the floor giving way and preventing him from escaping. It also involved lots of added CG debris and dust, sometimes interactive with the actors. The exterior, however, was pretty much entirely digital.”

“The biggest challenge was making sure that we were giving Chris Townsend and Marvel the sequence that they had imagined a year and a half earlier,” states Bryan Grill.  “It's the first big effects sequence in the film, and we wanted it to be everything they had hoped it to be and more. We had to remain flexible as edits and storylines sometimes do change but continue to execute and deliver photo real work in a comic book world.  I can’t help but think that the helicopter attack in Apocalypse Now [1979] was an inspiration.”  The visual effects contribution was a global effort.  “We want to thank Christopher Townsend, Victoria Alonso, Mark Soper, and everyone at Marvel for giving us an opportunity to work on such an amazing tent pole visual effects movie.  I think our teams at Scanline Los Angeles, Vancouver and Munich should get a shout out, too. They all had a very important role and should feel proud of our overall contribution to Iron Man 3.”

“We went to Trixter in Germany where I’ve gone several times before and their animation is phenomenally good and strong,” states Christopher Townsend.  “We needed a company there that could work with us and be flexible and that’s what they offered to us.  When we first started the project we literally had Scanline and Digital Domain and then Trixter.   We brought them on to do what we called the Suit Connect Sequence; it was something we needed for Comic-Con which was happening five weeks after we had stopped principle photography.  We worked with Trixter for several weeks before starting principle photography, designing individual suit pieces and the way that they moved.  Their concept artists started showing us some designs.   We found the things we liked, and they began doing modelling and animation tests.  We shot this complex scene with Robert standing in his suit in the garage receiving individual pieces.  We would project the pieces onto his costume and photographed with him the main camera but also with two or three witness cameras which were Canon 1600s; we used them to help track the body. We had Robert on wires flying off.  It is a sequence similar to when we see Tony put his suit on in the first Iron Man [2008].  It was homage to that suit up scene but it is now more several steps furthered advanced. Trixter was able to animate the pieces and have them attach onto him.  We tracked those pieces; it was great to have them do that.  They did several other sequences where there are much more subtle performances.  Iron Man is sitting on the couch or standing up or working across a room and giving a massage.  Its day to day things you wouldn’t expect Iron Man to do.  The subtle animation and the nuances they brought to sequences were great.  They also did a final sequence where Tony is dressed in just a glove and boot, and is fighting the bad guys and is being propelled.”

“It was a collaboration,” states Trixter Animation Supervisor Simone Kraus who worked with colleague Alessandro Cioffi on Iron Man 3.  “We’ve been doing projects together for six years or so, we’re a good team, and as a visual effects supervisor  Alessandro is  responsible for everything overall, but he has more focus on final picture. We worked together throughout process.”  Cioffi is well familiar with the big screen projects of the Marvel Universe in particular Captain America: The First Avenger and The Avengers.  “This time was different because we were working entire sequences.  We have three main sequences done from A to Z.  Plus a number of shots, other little sequences plus shots for both trailers and other sequences in-between.”  Over 200 visual effects were produced for Christopher Townsend who has frequently worked with the company headquartered in Germany since handling the European vendors for Ninja Assassin (2009).  “When you first start working with a production visual effects supervisor it takes a little time to adjust but then that collaboration consolidated in a strong understanding so I know his visual wishes.  Just to give you an easy example.  When starting on Iron Man 3 in April 2012 we went through extremely long cineSyncs sessions as we had to establish a number of facts happening in the first sequence.   But by the end of the show, I’m talking March 2013, we barely had any cineSyncs, we were talking on the phone sometimes, exchanging a few emails, and reading notes.”

A major reason why Trixter was chosen by Townsend was its ability to mange the ever changing needs of the production.  “It helps being smaller and knowing your people well as it makes things efficient,” notes Simone Kraus.  “We tried having experts for each individual thing. Pick the best individual, one person great at inner suit mechanics, and so forth, having Head of Art Department sitting next to modeller and making sure visual language was the same, that helps being efficient. In general we used the time before we had shot material to establish a visual language and get it approved by studio, in terms of Concept Animation.”  Alessandro Cioffi concurs with Kraus. “This is one of the few weapons that we have to use.  We need to stay efficient.  That means rationalize and optimize our resources, finding the solutions in the briefest time possible so we can provide the client with the visual solutions right away and then in the background work to finesse the solutions and present them as final a few weeks later.”  The workflow had to be partially altered.  “We’ve been doing massive animations for other types of shows so we have complex animation and rendering pipelines.  Our pipeline got modifications along the time we worked on Iron Man 3.  For example we introduced the use of Katana, this was about 50 per cent of completion on Iron Man; this was an important and massive change in our rendering pipeline.  We switched to the latest version of RenderMan so we did some important changes in our workflow which made things easier.  But for the rest we have a stable and complex animation pipeline because we were engaged in several local and international shows featuring a lot of animation.”

“The Suit Connecting Sequence has a production history that I wish to tell as this was the first sequence we worked on,” remarks Alessandro Cioffi.  “This sequence was selected by Marvel to be presented at Comic-Con in San Diego in July 2012. When we got the job we started working on it in the beginning of April; this was fantastic news for us that we had got to put our hands on a celebrated and iconic sequence but at same time it was a time race.  We had to literally crank up the multiple assets in no time.  All we got was the brand new design of the Iron Man Mark 42 suit from there we had to conceptualize and design all of the parts of the suit disconnecting.  We had to slice up the main suit into many parts and components.  Tony Stark does not need any machinery like in the previous movies as every single element flies off and connects to his body.  The first challenge was designing this untransformed element because one of the wishes from Marvel was, ‘We don’t want to see that this piece is a hand or a boot or a chest.’  They need to look as an undefined piece of technology with the same colours of [the suit].  The first few weeks were spent thinking about how these parts would look like and how they could transform while flying into the final configuration.  We had to create some animation that could be faithful and coherent to the previous Iron Man which we all know.  We had to make the technology look even slicker and new.  The first thing we did as a team was to go to our projecting room and watched all of the movies over and over to get visual references.  We realized in any connect sequence before Tony Stark stands still somewhere and either some machinery assembles the suit around him or in The Avengers he’s falling down the skyscraper in a still position and the suit connects onto him.  The gag in the sequence is that Tony is experimenting with this and therefore things go slightly wrong so the pieces are flying either too fast or wildly and he gets hit by every single part.  Robert Downing Jr. was great performing the impacts and was imagining what part was connecting to his body at a specific moment so his actions were frantic and his body was moving a lot.  One complication was to match move every part of his flexible human body in a way that would be compatible with armour which is something that’s rigid.  This was quite a complex process I must say.  We came up with some rigging solutions that made it possible.  The big reward in this sequence is the final animation when Tony does this summersault when he has to connect the mask onto his face.  The character is full CG and Robert Downing Jr.’s face is projected onto geometry and then he lands on this three point landing position which is an iconic image of Iron Man.”

The illusion of weight and flexibility had to be incorporated into the digital animation.  “We had to go through an insane number of rigging versions just to achieve that,” reveals Alessandro Cioffi.  “There was no way of using a single asset or rig and work around it.  Every single part of the sequence had its own specific issues that we had to work separately.  You can imagine a human body is extremely flexible and Robert Downing Jr. was contorting his body, imagining being hit by heavy metal elements.  We had to accommodate our animation to these types of movements.  The tight deadline forced us to start imagining how the connecting animation could have looked like way before we received the actual plates.  For this reason we were working on this animation language around turntables.  We designing and imagining every single bit of the suit connecting on a 360 degrees level.  We weren’t hiding anything conveniently like sometimes if you have a camera it might happen.  Sometimes you want to focus on the angle you’re working on.  Instead of waiting for the plates to be ready because they had to shoot it.  They shot it in May and partially in June.   We got started animating things and presenting them on turntables.  This connecting animation was then mapped onto the match moving rig that we had created in order to match move his body.  An entire amount of work was already there for us to play with for the final animation.  This was quite an advantage because we reused the animations because they were complete.”

"We had an early model [of the latest Iron Man suite], I think Digital Domain’s version,” remembers Simone Kraus.  “We started modifying it for our purposes. We had to deal with it in clean and destroyed versions, and in pieces inside/outside which were damaged or clean.”  The physical have suit created by Legacy Effects was a useful point of reference when establishing the digital copy.  “We have in this movie the Mark 42 looks slightly different in the material than the previous Marks in the other movies,” explains Alessandro Cioffi.  “This time we had some materials which were called the typical red which is more Burgoyne this time, a bit darker.  The double-coated red part was slightly less reflective than the suits we knew before. There was a metal quality which is non-reflective but is a sparkly like Champagne.  It’s an interesting material plus a more well-known brushed gold and dark metal; these two materials the red and Champaign that we had to come up with shaders right at the beginning.  Of course, if you look at the Comic-Con version they look quite different from the final version because we kept working on the shaders later on.”  Cioffi observes, “Sometimes when we talk about a digital character we forget that we start from a solid base which is how the materials get rendered without which you can never look right.  A background achievement which gave us a solid base to move from has been getting the shaders right.  This is something I think has been absolutely crucial for having successful work on this movie.”

“The Glove and Boot Fight is an effects oriented sequence,” remarks Alessandro Cioffi.  “It is spectacular because this is a showdown fight between Tony Stark who is only wearing a boot and glove, and a number of foes in the Mandarin’s headquarters.  There are a lot of effects, smoke trails given by arching boosters on the glove and boot because he is flying around plus some explosions here and there.  It is a visually and spectacular sequence.  The interesting part about of this were the suit components, in this case the glove and boot are heavily damaged.  We had to use completely different assets for this as we were changing the whole set of textures.  We were providing some heavy modifications on the assets to convey the idea that they had been heavily damaged.  We worked mainly on Maya and Katana for lighting, and rendering with RenderMan but for the smoke effects and all the others effects like explosions and fire, we preferred going for a plug-in of 3D Studio Max called FumeFX but rendered with Vray.  We had in the beginning these two parallel pipelines which merged together into Katana via an Alembic cache importer so we could join them Katana and render them separating, one with Vray and the other with RenderMan.  This was an interesting experiment which was successful because they blended well together.” 

“At one point there was a complicated stunt which Tony Stark had to perform which was done by a stuntman and this is a summersault across the room where the fight happens,” says Alessandro Cioffi.  “The reason why this shot is different is because the stunt is so complex they couldn’t shoot it on the location.  They had to shoot it separately in a green box.  We had to reconstruct the entire environment around the green box around the action.  This is the only shot which has this peculiarity.  We had to rebuild the entire location via projecting tiles shot on-set at 360 degrees.  We had an insane amount of tiles to reconstruct the environment.  We treated it as full CG shot as even the foreground and the stunt was almost completely rebuilt as a digital double.   The director and Chris preferred to change slightly the action so his limbs are digital and besides that we had to replace the head of the stuntman with Robert Downing Jr.’s digital face so at the end of the day it resulted in an 85 to 90 per cent full CG.”  Producing the smoke trails was not an easy matter.  “In the beginning this was the result of an insane number of iterations, and tries and failures because it is a subtle balance that you need to find amongst an insane amount of parameters.  The detail of the structure, the speed, and the way it interacts with the environment is crucial to belief it.  We had quite a long time where we were experimenting a lot until we found the formula to make it look right so from that point of view we could crank out shots one after another.  The plug-ins we used for the smoke effects were a fantastic tool.  For this specific type of smoke like propelled exhaust it has been a great help.”

“I’m sure you’ve seen stills of Iron Man on sofa,” states Simone Kraus. “Pepper [Gwyneth Paltrow] comes home and wants him out of suit. Its’ very romantic, and we see him do things we haven’t seen before, normally he’s attacking/fighting. But they almost kiss, come together, and feel close. Tony Stark massages her shoulders in the suit; the whole thing was about making the audience forget he’s in a metal suit where you can’t do subtle movements.” The uniqueness of the domestic sequence caused some complications for Alessandro Cioffi and his team. “The main question for us was, ‘How does Iron Man behave in such a situation? How does he move? How does he walk? How does he hold his hands? How does he look like? It might sound like trivial questions but there is a lack of references in the previous movies. We never saw Iron Man walking in his living room. We never saw Iron Man having a tender moment with his girlfriend or trying to kiss. How does Iron Man give a gentle massage to his girlfriend? Besides of the technical problems of having interaction between the CG character and an A-Class star like Gwyneth Paltrow [Se7en] there were the questions, ‘How does he behave? What’s the style?’ We could only partially refer to Robert Downing Jr. as an actor in his performance as Tony Stark because we didn’t have any direct references. And besides the Iron Man wouldn’t move like a man anyway; he is still a man in heavy armour. We went from animations like he is cool guy quite macho walking around in his living room and then we went he is more humble, tender, gentle and decent. We tried many solutions until we found one that was well received both by the director and the visual effects supervisor; that was quite a challenge. Iron Man gets to perform some funny gags. There is one shot in particular where he is looking quite quizzical at Gwyneth Paltrow. I have that very frame. It’s really funny.”

“The greatest challenge was time pressure for Comic-Con, doing everything at same time, keeping an overview and getting it done in time,” states Simone Kraus.  Suit Connect is interesting because its detail oriented from a technical point of view. You see inside and believe it even though we made it up.  I like the Date Night Sequence that we were asked to do something we haven’t seen before, a date. Glove/Boot was cool because of the great action, reminded me of IM1, when he’s trying to fly around and not damage his own car.  I don’t have a favourite.  I have favourite things in all of the sequences.   Great experience overall.”  Kraus adds, “The show was perfect for us because we could use our Art Department, asked to develop suit, normally someone else would take it and do something else. But this time we did everything, design, model, rig. Best use of Trixter’s creativity in a show.”  Alessandro Cioffi believes, “Every part of the movie has been a different challenge.  What I’m particularly proud of in that sequence [Glove and Boot Fight] are the effects.  The effects are believable.  You don’t look at them like a digital intervention.”  A signature cinematic moment resulted in some doubts.  “There are a couple of shots like this I’m particularly happy about because I thought, ‘Oh, my God this is never going to turn out right.’  Then we were happy with.  One is definitely the iconic three point landing pose that is at the end of the Suit Connecting Sequence.  That one is like the golden boy.  There is another one in the Glove and Boot Fight which had been tricky but eventually turned out right.  Robert Downing Jr. was heavily rigged but had to perform like the boot is propelling him up; he is flying around the room shooting around with a machine gun and blasting with the RT from the glove around so it is heavily dense with effects. That shot for some reason didn’t come right at the first go so we had to work in every department from removing the rig as he was wired, extending the set, adding the effects, rendering the technology parts like the boot and glove. This shot in particular I cannot describe it better than I did so far but it is one of those when I look at I say, ‘This has been quite something and I’m quite happy about.’  Cioffi concludes, “I always describe this job on Iron Man 3 like a bumpy take off, an enjoyable flight and a smooth landing.”

Method Studios was recruited to handle 80 visual effects shots over a three month period.  “I had worked with them before and enjoyed that experience and was pleased to be able to work with them again.  Method did the Water Tower Collapse Sequence and a couple of sequences with Iron Patriot.”  Leading the facility team in Los Angeles was Method Studios Visual Effects Supervisor Matt Dessero.  “The interesting thing creative freedom wise we had to create a couple CG shots, one of them was completely 100 per cent synthetic.  It was all based on photography and this was the Rose Hill Tower Sequence with the water tower destruction shot.  There was a LiDAR of the entire scene.  We were grounded into what was captured and what plates were shot on.  The guys and Chris Townsend had been working on the Extremis Effect for a few months before us getting on the show.  Chris was forthcoming with sending visual references over to us and geometry in some cases.  We had a vascular system which was one of the key components that helped to conclude the effect with the internal glow.  We did a plant regeneration shot where Chris said, ‘We need this plant to regenerate.’  He gave us a few notes, ‘It would be good to go from here to there.’  We had full freedom on that to design and develop what we needed.”

“In the Rex Sequence we had not only the Extremis Effect but also had the Iron Patriot suit opening,” states Matt Dessero.  “The look had been established at this point by some of the other companies when we got onto the project later but we still had the challenge of bringing it to a hot point where we added a blistering effect to the skin.  Each shot started with a tight camera track. Once the camera tracks, model and rig were approved we began match moving the actor’s performance.  To further lock the CG model to the plate our modellers built custom blend shapes. On average there were three blend shapes needed per shot to lock the actor’s facial performance. The blend shapes were modelled to match key points in the actor’s performance. These match moves were important because the Extremis Effect not only has a surface component, but the richness Chris Townsend was after comes from the deeper internal component of the effect that included an internal glow that is occluded by the bones, and vascular system.”  The next challenge was to create the opening of the armour suit. “The Iron Patriot model has a lot of moving parts for both the exterior and interior. Rigging these pieces required a bit of back and forth between the modelling, rigging and animation departments. Due to the limited time we had to complete this effect we only built what was necessary for our shots. Animation began with a camera track and match move of Rhodes’ [played by Don Cheadle] performance. We roughed in the timing with very few simple exterior shapes for the helmet, chest, arms and legs. Getting the opening to flow and feel purposeful was important. Each animated event was motivated by Rhodes’ performance. Making sure the Iron Patriot felt heavy was Chris Townsend’s primary concern; he never wanted the suit to feel flimsy. This heft was achieved in many ways, some of which included adding reverberation which was added as each large piece opened; we also stiffened the joints to limit arm and leg sway. Once the primary animation was approved, we scoured more of the suit, added more movable geometry to both the interior and exterior, thus adding complexity to the suit opening effect.”

“The Rose Hills Tower location was LiDAR so we had full geometry data that matched to build our water tower,” explains Matt Dessero.   “Tony Stark is without his full Iron Man gear; he has his hand and chest RPs and that’s about it.  Savin has learned about the Extremis Effect at this point of the show; he puts his hand on one of the legs of the water tower and it starts heating up and turns to molten metal in one of our shots.  The water tower collapses under its own weight. Chris shot a bunch of dump tanks water shots. They’re beautiful photography.  The challenges here were the molten leg which we rendered a fluid simulation.  We not only have that but the leg heating up, ashes, embers, paint flecks spewing around and snow melting off the water tower in a  the snowing environment.  As he pulls the leg down the water tower starts to sway back and forth, and snow clumps start to drop off a bit.  We go to an overhead.  The tower is collapsing and water is pouring out of the water main and out of the bottom of the tank.  The tower falls, hits the ground, collapses and at that point we cut to a bunch of shots that Chris shot of the water tank.  They are more medium wide type shots.  We built the wide shots.  The wide establishing shot they shot a crane plate and we added our water tower collapsing into that.  The camera goes down low to the ground as the water tower collapses.   There is water rushing towards the camera which we added a bunch of CG water splashes and trees getting blown around from the water rushing through.”

“There was an actual water tower on location that they shot around,” states Matt Dessero.  “We painted out that tower and added our CG tower in because our tower needed to sway and move around.  But for the low angle shot we didn’t have a plate we had to create that from scratch.  There is water rushing right towards and underneath the camera so it would have been a tough one to capture but for all of the medium shots where Tony Stark is interacting Chris shot the dump tank and there is also a portable construction trailer that gets knocked. What was beautiful was that Chris shot a giant dump tank there and pulled apart the trailer.  We always had something grounded in the sequence. The majority of the shots are all practical but CG was needed for the shots where the water tower is swaying, and water is rushing right at the camera.”  To realistically produce water digitally a number of layers were required.  “There are foam, froth and spray, and bubbles underneath the water,” says Dessero.   “We added atmospherics, a fine mist to help make it feel dense and photo-real as well.  There is layer upon layer created from a simulation and it is also getting the timing right.”  The centrepiece of the sequence was helped by a practical plate.  “The water tower was texture painted in Mari.  It couldn’t be too stained.  It was white paint which is always a fun one.”  YouTube led to some creative interpretations.  “We looked for a lot of reference for water towers collapsing and they’re actually boring.  The water towers when they’re collapsing they’re usually being knocked over by natural forces or knocked down for safety reasons.  They just fall which is pretty uneventful.  We did find some that had thicker gage steel walls and some that had thinner gage steel walls.  We went somewhere in the middle.  We didn’t want the tower to feel too flimsy or like a steamliner.”

“I like our plant rejuvenation shot,” remarks Matt Dessero.  “Happy Hogan who is Jon Favreau’s character rips a piece of a plant off and walks away.  We cut back to the limb of the plant re-growing, leaves start to unfurl off of it and there are ashes and embers falling off.  For reference we looked at some time-lapse for plants.  It’s not as jerky as the time-lapse because things move in the wind and the weight of the plant drops so it is a smooth continuously flowing shot.  There was a plant that they shot so we removed a couple of limbs to clear it out for composition and then we added our CG limbs. When we got into the design we painted out a couple of designs the way we thought it should look.”  The concept for the moving organic object was not presented to Christopher Townsend and Marvel in the completed stated.   “We started with a low resolution.  Once the simulation is brought into grey scale form we get into texturing and rendering. We started off with grey scale and got everything approved animation wise.”  A 2D approach was adopted.   “I personally like to resolve as much as we can in 2D concepts before we get into building it in 3D which is slower, costs more money and takes more time.” Extra detail was incorporate into the images. “There were some nice bits we added design wise like the burning blistering or the rejuvenation or the extra water detail and the molten metal.”  Dessero states, “We didn’t have a lot of time and there was a lot of technical challenges and at the end of the day some beautiful work.”