Reallusion (R): Hello Cesar, please tell us something about yourself.

Cesar (C): Hello, let me start at the very beginning. i started my journey as a filmmaker in 2004 with my first war film about the 1982 Falklands War, for which I produced VFX close-ups that seamlessly blended real images with 3D animation. I went on to direct my first documentary film on the same subject for the History Channel, which featured over 60 minutes of captivating aerial combat animation.

As part of my creative journey of discovery, I ventured into the area of fan films and gained worldwide attention with a production about Robotech. In Japan in particular, the film received a huge response with over 200,000 views within one day.

R: How did you come up with the idea of making the science fiction film “Nick 2040”?

C: Based on my previous experiences, I ventured into my own science fiction world, which I called “Invasion 2040”. This transformative endeavour took seven years. Along the way, I secured an Epic MegaGrant and our project caught the attention of Reallusion, whereupon we participated in their Pitch & Produce programme.

With “Nick 2040” I went one step further and raised the production to cinema level, I filmed with RED cameras in 6K. The post-production was meticulously rendered in 4K, completing the trilogy of short films. This last part serves as a proof of concept for an ambitious feature-length film, which is currently in the pre-production phase.

The prerequisite for this is that a large part of the VFX preparations have already been completed before shooting begins. At the moment I am actively involved in this preparation phase – every scene, character, animation and VFX test created for the short film will be seamlessly integrated into the full feature film. This dual role allows me to act as both co-director and VFX director.

R: Please describe your workflow using the Reallusion software?

C: Phase 1 – Creating realistic characters with headshot and dynamic wrinkles

I started by scanning the robot’s face with Reality Scan. At the same time, I created a face from a photo using Headshot. In parallel, I used ZBrush to connect the two models and create a mesh that closely resembled the target face. I then imported the mesh and texture created in iClone into the Character Creator using the Headshot plugin. Once completed, I transferred all the components into the Unreal Engine to generate the robot’s face in a metahuman framework. Additionally, I scanned the clothing to improve the realism of the final rendering.

I carefully worked out the antagonist, a tall giant, and designed the facial features down to the smallest detail. I was inspired by the facial expressions of the metahumans and went beyond conventional methods. Using Blender, I seamlessly merged an AccuRIG body with a metahuman face, which led to an amazingly impressive result.

The key breakthrough was creating facial animations with Dynamic Wrinkles. This allowed us to realistically render the protagonist’s face as well as the characters in the invasion scenes.

For some scenes, the actress’s face had to be recreated, which was not possible with conventional shooting methods. I scanned the actress’s face with the Headshot plugin and optimised the mesh with ZBrush in combination with CC4. I then animated it in iClone 8 and seamlessly transferred all the data into the Unreal Engine via Live Link, which led to a fantastic result.

With CC4’s advanced dynamic wrinkle features and the use of face mocap, I was able to create complicated facial expressions effortlessly in a matter of minutes.

Phase 2 – Animating the fight scene with iClone and ActorCore

The integration of iClone played a crucial role in the development of the transformation animation for Invasion 2040. This was particularly crucial as the character seamlessly stands up and places his hands on the ground as he undergoes a dynamic metamorphosis into a robotic form.

The invasion scene in “Nick 2040” was carefully crafted in the Unreal Engine with a mix of characters from different platforms – some from Metahumans and others from iClone. Using the handy and easy-to-edit Actorcore scans retrieved via iClone’s Live Link proved ideal for seamlessly integrating the various animations from the“Run For Your Life” pack and perfectly complementing the intensity of the invasion sequences. To create a cohesive scene, I cloned the iClone animations onto the skeletons of the Metahuman characters in Unreal.

Reallusion and iClone 8 allowed me to create complex animated scenes with epic battles between giants and robots. These scenes, which are at the heart of my films, are brought to life by the powerful features in my workflow pipeline. The possibilities offered by these tools, especially in the field of computer graphics and animation, play a crucial role in realising the vision and complexity of my cinematic creations.

Phase 3 – Immediate preview of results with Unreal Live Link

The Live Link extension is truly remarkable. The process of transferring animations to Unreal is both seamless and stable. Correcting a motion is a simple task – I make the adjustments in iClone and make sure I adhere to the start and end times on the timeline. Then I simply replace the animation with the Unreal Sequencer and complete the process with a few clicks. In a few moments, the updated animation is finished and ready to render.

Timecode Sync Rendering has also revolutionised my workflow. Now I can seamlessly transfer objects into iClone, animate characters and use these objects as references. The speed at which I can send the animation to Unreal is remarkable. It only takes a few seconds. This streamlined process has made animating incredibly easy and fast, which has greatly improved my efficiency and creative ability.

The final fight between the robot and the jumper was a two-step process. Firstly, I choreographed the movements of the two characters using Mocap, executed the choreography in person and perfected the hits in iClone to achieve effective results. These animations were then sent via Live Link to Unreal for rendering. At the same time, the fight scene was supplemented with additional sequences from the Reallusion Marketplace to increase the dynamics and intensity of the fight scene.

R: Why did you choose Reallusion?

C: We are currently finalising the storyboard, shooting, budget and VFX plan for “Nick 2040” – an important task that forms the foundation of our feature film. In a compact animation studio where the tasks are inherently complex, our small animation team has developed unimagined possibilities through this process.

• Stability: The entire process is characterised by maximum stability without crashes.
• Speed: Animations with more than 1000 frames are transferred to Unreal in less than a minute. It is remarkable that the processor and GPU are often not overloaded during this fast process.
• Organisation: I streamline my workflow by storing different animations in iClone’s Content Browser. This allows for quick retrieval and use in different scenes of the film, as virtually all the sequences I need are stored there.
• Practicality: Integrating animations from various sources, including Noitom Mocap, Mixamo animations and Unreal character animations, is a breeze. Simply drag and drop these animations from Explorer into iClone and they are seamlessly integrated into the project.

Nearing completion, the trailer gives a first glimpse of our work and we look forward to the release of the final short film.

Learn more

• Episode Nick 2040 created by Cesar Turturro
• Animating 3D characters and film scenes with iClone
• Create realistic characters with the Character Creator

The post Mech vs. Giant Slugfest mit iClone & Character Creator first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

In nuce: Mark Robert Motion Control (MRMC) has announced that it will be present at this year’s IBC (International Broadcast Convention) with a selection of motion-controlled robots – including: related control software and broadcasting techniques. MRMC Chairman, Assaff Rawner, expressed his confidence in MRMC’s presence at IBC. MRMC will be on stand C30 in hall 11 in Amsterdam.

What can visitors expect? From 9 September to 12 September inclusive, visitors to the MRMC stand in Amsterdam will be able to experience the Bolt Jr Cinebot performing its automated camera movements. Visitors will also be able to take a look at a demonstration of the Flair software – among other things: The most important functions of Flair and the newly added Focus Assist tool. The following devices will also be presented: the ARC-360 PTZ camera, the QRS-1 Quiet Rail System, the StudioBot XL robotic arm and the TRH-1 rail system. According to the manufacturer, the QRS-1 Quiet Rail System is both an affordable and customisable rail system, while the StudioBot XL robotic arm was designed for automatic camera movements for news and studio recordings. MRMC will also be present with solutions that are especially suitable for the broadcasting of sporting events. In its self-presentation, MRMC states that all of the manufacturer’s robots are camera-independent and work in combination with any camera model.

Click further: To the Mark Robert Motion Control homepage and the IBC website. Digital Production last reported on IBC 2022 on 26 August this year , when it was Maxon’s 3D and Motion Design Show that advertised its presence at the trade fair this year – with many experts sharing their expertise on Maxon tools. We would like to thank our colleagues from Provideo Coalition, from whom we obtained some information for this report.

The post At the IBC 2022: MRMC with motion control robots first appeared on DIGITAL PRODUCTION and was written by Patrick Poti.

Industrial Light and Magic, the world’s biggest effects company, raises the bar even further in many ways to realise giant robots and other effects for Michael Bay’s “Transformers: Revenge”.

It’s bigger than the first film. Much bigger. Huge, to be precise. Enormous, colossal. “We used 95 per cent of our studio’s render farm,” says Scott Farrar, Visual Effects Supervisor at Industrial Light and Magic and responsible for the special effects on “Transformers: Revenge of the Fallen”. The original film title of the Transformers sequel is “Transformers: Revenge of the Fallen”, directed once again by Michael Bay.

The DreamWorks film once again sends the good Autobot robots into battle against the evil Decepticons. The main protagonists on the human side are once again Shia LaBeouf as Sam Witwicky and Megan Fox as his girlfriend Mikaela Banes. The Autobots Optimus Prime and Bumblebee are also back, as is the Decepticon Megatron.

This time, the battle moves out of Los Angeles and takes place in a total of seven states and three countries. The film also takes place in Egypt and Jordan. This time, 45 new robots join the two opposing parties and the other robots from the first film.

“We needed between 16 and 20 TBytes of hard drive space to render the first film,” says Farrar. “This time it’s 140 TBytes.”

ILM created 560 shots for the film, including the 60 robots. Some of these shots were 500 frames long and showed scenes with three to four large robots. This is complicated and extremely time-consuming when it comes to rendering. In addition, Digital Domain created some robots: the Decepticons Reedman and the Microcons Soundwave, Wheelie, the Kitchenbots and Alice, a hypocritical robot.

To assist in the creation of the 60 robots, ILM created a digital model shop where modellers could create a library of robot parts. Anyone creating a character could then use any of these parts, which already had a detailed texture map. If a part of the library was changed, every robot that used this component also changed automatically. The modellers created the characters using Maya and ILM’s own software Zeno.

Animation Director Scott Benza headed a team of 50 animators at ILM. 20 of them were already involved in the first part of Transformers. The large bots, consisting of many parts, were almost always animated by hand. Rigs from the first instalment were also used again, which made it possible to combine the components as desired.

“If someone was working on a shot and came across a new hole in one of the robots, they could simply take a part from another robot and fill the gap,” says Jeff White, second visual effects supervisor.

For example, during an emotional scene with Sam (Shia LaBeouf), a flood of tears pours from Bumblebee’s eyes. However, this robot model was not designed to allow tears to flow. But because parts could be combined, an animator took some parts, added windscreen wipers and used the windscreen wipers’ spray system to create the tears.

Of the 60 characters ILM created, 45 were new heroes and 40 had at least one line of text and had to be animated accordingly. A robot over 30 metres tall called Devastator, made up of seven other robots, was the largest in scale.

Also new is a female Autobot, Arcee, who is built from three motorbikes or “Doc”, a small Decepticon who transforms into a microscope. Jetfire is an older robot that walks on a stick and was originally a Decepticon. That Jetfire transforms from an SR71 Blackbird, a jet that was used as a spy plane in the 60s. It is over 12 metres tall and built from old mechanical components. Even though Bumblebee cries in the aforementioned scene, the aged Jetfire has probably the most emotional and human traits of all the non-human protagonists.

“They breathe, spit, spray sparks and produce gas and smoke,” says White about the robots in the film. “We wouldn’t have thought of doing that in the first part of the film.”

As well as making the change to give the robots more personality and emotion, Michael Bay also took away their metallic gloves, enhancing the fight sequences and making them more violent. This culminates in Bumblebee tearing another robot to pieces in one scene.

Scott Benza was responsible for putting together the team of animators. Some of them were hired specifically for this project based on their skills and interests. Those who had great skills in animating animals worked on Ravage, a cat-like Decepticon. Tech-savvy animators took care of the transformation shots. Those working on Devastator were given more powerful computers. The giant Decepticon required 16 GB of RAM. Formed from seven large vehicles in robot form, the enormous machinery consists of 52,632 parts and around 13 million polygons.

In order to be able to work with the large robots and the large number of individual parts, ILM created a production pipeline with different resolutions. The animators had many buttons on their screen that allowed them to change the resolution from 25K or lower up to 1300K per frame. Each individual part of a robot could be selected and viewed and edited in the desired resolution. At the lowest resolution, even Devastator could be moved in real time.

Devastator resembles a gorilla with an open stomach surrounded by tooth-like components. He is created by brutally assembling himself from one vehicle after another, starting with a mining excavator. One of his arms is made from a crane, the other from a kind of shovel excavator. His legs were originally a rubbish lorry and a bulldozer. A second rubbish truck is used as his torso and a cement mixer becomes his head, into which he sucks everything that gets in his way using a vacuum. “There is one shot in which Devastator moves towards the camera. If you wanted to render that one scene on a home PC, it would take three years in full resolution,” says Jason Smith, Digital Production Supervisor.

Michael Bay worked directly with the animators at ILM to create camera movements for the animation sequences and to contribute his own ideas. “He was constantly on camera. Whether on set or at ILM,” says Benza. At ILM, Bay controlled a virtual camera on the studio’s motion capture stage to film the animated characters. As on set, Bay improvised with the actors. But at ILM, it was animated characters and the animators. Sometimes the animators had ideas that Bay used for the film. And, after ILM choreographed a huge fight between Optimus Prime and Megatron in a forest, Bay invited Benza and Farrar directly to the location, a forest in New Mexico, to supervise the plate photography with their Previz.

For this sequence, the effects crew created CG trees in post-production to adapt the shots to the location. This allowed the robots to crash into the trees and break branches. Each branch was a collection of simple objects strung together. The technical directors were able to run a rigid-body simulation, which resulted in a high-resolution model.

Size matters

With three times the number of robots, which are also significantly more complex, ILM placed very high demands on the available mass storage. In addition, some scenes were shot in IMAX resolution. The studio normally renders the scenes in 2K resolution, but the IMAX scene has eight times the number of pixels at 4K resolution. “We had more compositing work in 4K,” says Smith. “The compositing artists spent a lot of time zooming within individual frames.”

The compositing for the IMAX shots took place in Nuke and Bay cut out the scenes and used them for normal projection. “You definitely want to see the film in IMAX,” adds White.

IMAX also had an influence on the effects crew. In the past, for example, sandstorms were created using instanced particle-based sprites. With 4K, the crew had to rely on volumetric simulations to create enough detail for the sand that the Devastator sucks into its cement mixer mouth.

In one notable scene, Devastator climbs up a pyramid in Egypt and knocks the top off as if it were an unloved toy covering an important piece of technology inside. As he does so, individual blocks can be seen crumbling off the sides and falling downwards. To make this shot possible, ILM’s artists use fields and particles to define fracture lines in 3D on a CG pyramid. The simulation team then ran a rigid-body simulation with 150,000 individual parts on GPUs instead of CPUs.

To create the dusty and shifting sand falling from the pyramid, the TDs built a grey cone and put a piece of cloth over it with a matching surface simulation. They then used the fabric simulation to create a particle simulation. Once they had approved the shape of the simulation, they ran the particle simulation.

It was ILM’s turn

There are two scenes in the film that were created entirely digitally. One takes place on Nemesis, the planet where the Decepticons come from. The other takes place underwater when the Constructicons and the DocBot rescue Megatron.

“That’s where we were needed,” says White. Bay again used the virtual camera to show the crew the framing he envisioned. Then it was up to ILM to reproduce the atmosphere he was used to on set. “He uses loads of drops, smears and fog,” adds White. “So they added all that to shots with The Fallen.” The Fallen is a creature that lies flat on its mechanical back and looks like it could be the Alien’s cousin.

For the underwater scene, the TDs paid special attention to how the lighting works underwater. For example, dirt was built in to create realistic reflections of the light beams.

“There are key scenes in every film,” says White. “In this film, every sequence has key scenes.”

Digital environments

In addition to the robots, ILM created digital environments. “We needed big backdrops for our big robots,” says Farrar. Although Bay filmed in Egypt, Jordan and other locations over a six-month period, the Digimatte department still had its hands full. In some locations, it would have been too difficult or too expensive to bring all the equipment for a particular shot. So it was a good solution for the digital artists to recreate locations using projection maps on low-res 3D geometry. Bay was able to “fly” through these virtual sets with his virtual camera and this became the most efficient way to create the shots. “That’s a good offer to make to the director,” says Farrar.

Even in places where Bay could film, the crew took enough digital photos to be able to recreate them digitally. That was in case Bay wanted to change anything in post-production. Which he did from time to time in collaboration with the effects crew.

“Working with Michael is always a lot of fun,” says Benza. “His films are the hardest but also the most enjoyable.”

And they are also the biggest and most brutal. Where else can you watch heavy robots made up of thousands of parts smash themselves to pieces? And if ILM has done its job well, the viewer will believe every minute of it.

The post Bigger & Better first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

No hair, no fur, no skin; robots actually guarantee the artists a quiet job. The production of Transformers was different: the highly complex extraterrestrial robots consist of countless individual parts. Added to this were difficult simulations for collapsing buildings. The team at ILM had to look for new approaches and came up with a number of solutions that made a decisive difference to the company and the industry.

Could there be a better subject for a summer blockbuster than watching gigantic robots from a distant galaxy smash themselves to pieces in the centre of Los Angeles? And can you think of a director better suited to staging such carnage than Michael Bay, who has already directed “Armageddon”, “The Rock”, “Pearl Harbor”, “Bad Boys” and “Bad Boys II”?

The names of the executive producer and the visual effects studio also stand for decades of experience with action films: Steven Spielberg and Industrial Light & Magic. George Lucas’ company took on the task of creating the massive bots and the enormous destruction.

Taken together, the mixture of Transformers, Paramount Pictures and Dreamworks SKG is this summer’s potential blockbuster.

From toy to film

Starting with the Hasbro toys that were launched on the American market in 1984, a small universe was created: in the 1980s, the Marvel publishing house first published a comic story, which was later developed into a television series and an animated film.

Transformers tells the story of the clash between two alien races, the Autobots and the Decepticons, who come from the planet Cybertron. Both are searching the Earth for a magical cube called the Allspark, which was once discovered by an ancestor of film star Sam Witwicky (Shia LaBeouf). Unbeknown to Sam, he possesses the map that shows the way to the hiding place of the Allspark. On Earth, the heroic and good Autobots disguise themselves by transforming into cars and lorries, while the evil Decepticons transform into military vehicles. The leader of the Autobots, Optimus Prime, is a lorry tractor, Megatron, leader of the Decepticons, is a Cybertronian jet.

New tools

To create the visual effects for this film, ILM had to develop new tools and techniques under the guidance of its Oscar-winning visual effects supervisor Scott Farrar. Selective ray tracing helped to reduce render times. A new type of rigging system and a controllable rigid body simulator helped the animators manipulate robots made up of thousands of individual parts. For the task of destroying buildings and concrete, ILM used both proprietary and off-the-shelf software.

“We call this a ‘hard body’ film, as opposed to productions with actors who have hair, fur and skin. So you’d think that shouldn’t be that hard,” says VFX supervisor Scott Farrar. “But we’ve never had to reach that level either. I worked on Minority Report, which was about concrete and cars, but we couldn’t work it out to the level we can now. We had to programme a huge amount of software for that film and solve countless individual problems so that the effects would actually look better in the end.”

Original vehicles and concept art

It starts with the robots themselves. Every single robot – except the Decepticon Scorponok – transforms into or out of a vehicle in the film, often in full motion. ILM had a fleet of real vehicles as reference material, including cars supplied by General Motors and military vehicles provided by the US armed forces – sometimes these were entire vehicles, but sometimes only photographs and videos were available.

For the robots, the artists had concept art as a reference. But when ILM received the concept art designs approved by the director for the film, it quickly became apparent that the concept artists had not bothered to build the robots from the parts of their camouflage vehicles.

“On some of the early robots, we were trying out how we could move a door into the chest piece and scale it,” recalls associate visual effects supervisor Russel Earl. “After a few weeks, we decided to take our chances, match the artwork and hope we could find a way to get them to transform.”

Therefore, unlike originally planned, the robots now also contain parts that do not actually belong to their camouflage vehicles. In some cases, this involves several thousand individual parts. “To achieve the mass that the actual shape of the robots requires, you have to use more than a car has in individual parts,” explains Earl.

Surfaces as if painted

A team of modellers led by Dave Fogler built the robots and CG vehicles using Subdivision Surfaces in Autodesk’s Maya. Viewpaint supervisor Ron Woodall led the texture painting team, whose job it was to add scratches, dirt and even cracks and tyre marks.

“A lot of stuff that you think is modelling is actually painted,” reveals Fogler. “We wanted to keep the models as light and economical as possible, but the debate isn’t settled on which is more complex to render: geometry or displacement. It could be a draw.”

Woodall adds: “For interactivity, displacement is the best method. But rendering geometry is cheaper.” The painters worked in Photoshop and ILM’s in-house software Zeno. The team quickly realised that the robots had to look dirty to be believable: shaders couldn’t do it. “Without the wonderful work of Ron Woodall, the technical directors would have been at a loss,” praises Hilmar Koch, who supervised the work of the TDs. “Synthetic shading would look boring without the painted effects.”

Selective ray tracing

The shaders had a different task. “I think one of our biggest achievements in terms of shading is that we can switch ray tracing on and off as required,” Koch reports. “With selective ray tracing, we can soften certain parts of a setting.”

In order for the robots to successfully and credibly fit into the respective environments and to make the interaction with the live actors more convincing, the targeted use of reflections was crucial.

“It’s difficult enough to light a car in CG so that it looks real. But we took the car, dismantled it into a few thousand individual parts and put it back together again in the shape of a robot,” Earl explains. “For example, if you take a part of the body and break it down into a bunch of individual parts, you’ve obviously lost the shape of the car, but we still had to give the audience the feeling that when they look at a particular part of a Transformer, they’re sure they know what part of the car it is. A lot of that effect is based on reflections.”

To reflect the real shot scene, the technical directors used 8K quality environment maps. They stitched these together from photographs that they had previously taken on location. Ray tracing was added as required for certain parts of the robots. ILM used Mental Ray for ray tracing on the one hand and RenderMan for scanline rendering and ray tracing on the other.

“We control this via the shaders and the user interface,” says Koch. “You can select an individual part of the robot, then click a button and the ray tracing starts.”

Part by part

Modellers built the robots from thousands of components. The smallest robot, the Decepticon Soundbyte, which disguises itself as a radio instead of a vehicle, has 871 parts. Optimus Prime, on the other hand, the leader of the Autobots, has 10,108 parts. To give the animators control over all these elements, the character developers developed a dynamic rigging system that allowed the animators to arbitrarily group and regroup geometry in different ways.

“What’s really unique is that the animators can move every single piece of geometry you see when a robot appears on screen,” praises Digital Production Supervisor Jeff White. “There’s a basic rig with animation controls that the animators use to move the robot around the scene. If they use a second level of refinement, they can group any combination of parts together and make them move completely differently. They put the pivot points exactly where they want them.”

Facial animation

This became particularly important for facial animation to repair intersections and interpenetrations of hard metal parts and create transformations. Many of the robots, especially the Autobots, have speaking roles. To create facial expressions, the modellers built a turbine system for the eyes with 50 metal arms that dilate and constrict like a pupil; for the lips, they used split metal pieces and fixed cones in the mouth to simulate the tongue and teeth.

Robots with body language

“Within a scene, we started our work with posture and body language. The next step was to try to convey a certain movement or feeling with the overall pose,” explains Shawn Kelly, who was one of five animation leads involved in the production. “Some robots are a little more athletic than others, but they’re all out there because they’re hardened fighting machines. They all have a similar fighting style. The facial expressions end up being the icing on the cake: The way their eyes move and where we place a wink.”

For the robots’ fighting style, the animators used a mix of video reference footage Michael Bay had shot of stunt actors miming Japanese and Hong Kong-style martial arts, as well as motion capture data from the stuntmen. “We used the reference as a guide, but we did a lot of the action with keyframes,” says Paul Kavanagh, one of the animation leads. “We always tried to give the impression that the robots have a weight of their own and can also move very quickly.”

Little tricks

The animators cheated with the transformations. “We used tricks,” admits Kavanagh. “We obviously tried to make the transformations look as real as possible and think of as many parts as possible, but we couldn’t always account for every detail. Once we had the most important parts working, we added the rest more or less on the fly and didn’t worry too much about it. What was a daunting prospect at first turned out to be a very enjoyable task to work on in the end.”

Generally, the animators animated the robot and its camouflage vehicle as they played their part and then handed over the poses to the creature development department for the transformations. However, as the work progressed and the animators used the dynamic rig, more and more animators also created the transformations.

Insect robots

One of the robots, the Scorponok, never transforms. Instead, its behaviour is more reminiscent of an insect than a human. Animation supervisor Scott Benza played the robot Scorponok directly, like a kind of virtual animatronic. He employed rigid body dynamics by using rotational forces, forward velocity and the like.

“We took our rigid body simulation to a whole new level for this film,” says Benza, looking back. “We now have methods that allow us to drive characters with virtual motors; we have settings that are driven by fixed simulations.” Benza baked the simulation data onto animation controls so that the animators could change them afterwards in the same way they would change motion capture data.

Massive destruction

Wherever the robots fight or simply drive down the road, they cause massive destruction. Michael Bay decided to shoot much of this destruction and explosions in real life on set, but for some shots, especially towards the end of the film, it was necessary to shoot them entirely in CG.

“We had to break up concrete and destroy buildings,” explains Koch. “We had bursting walls and falling debris. And we built quite a few smoke trails, all with particles. We set fires, knocked down lampposts, street signs and traffic lights, smashed windows, ripped fire escapes off buildings and destroyed all sorts of structures. Without 64-bit rendering and 64-bit computers, we wouldn’t have stood a chance in this production.” The team used both proprietary and off-the-shelf software for the destruction.

Difficult conditions for lighting

ILM’s lighting artists had to overcome a very special challenge in this production: The smoke-filled scenes with the giant robots were particularly complicated to light. To ensure that the lighting artists could work faster, ILM’s research and development department supported the team in improving the lighting tools in the studio’s own Zeno software.

“We tried to hide as much of the underlying infrastructure as possible,” explains Senior Technical Director Hilmar Koch, “so that the rendering TD only saw what really mattered.”

While most cinema-goers will probably be concentrating on the action in Transformers, visual effects supervisor Scott Farrar is delighted with a success of a more subtle kind. In his opinion, the modelling, painting, rendering and lighting in this film represent a turning point for the visual effects industry.

“We’ve always had the iron rule that, in principle, it looks better to shoot a model in real life than to realise it using computer graphics,” he explains. “But I think we’ve finally overcome that hurdle with the achievements of the production of Transformers.” And that’s a transformation that even a Cybertron would approve of – whether Autobot or Decepticon.

The post Thousands of parts first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

The Canadians from Monsters Aliens Robots Zombies VFX you: Shattering shotgun men, bullets ricocheting off children’s chests and Ellen Page’s feet – bathing in a tub of the dozen-eyed kind. There’s more supernatural stuff to watch in the VFX breakdown of the second season of The Umbrella Academy.

The post The Umbrella Academy | VFX-Breakdown first appeared on DIGITAL PRODUCTION and was written by Patrick Poti.

The latest breakdown from the Swedish VFX house with the tongue-in-cheek name Important Looking Pirates sees an android pulled out of the retort, the face of series heroine Dolores pushed together and skin slipped on like a glove.

The post Westworld Season 3 | VFX-Breakdown first appeared on DIGITAL PRODUCTION and was written by Patrick Poti.

In the latest VFX Breakdown by the Rodeo FX Company, robots stomp through the scenery, buildings are dismantled and prosthetic arms are attached. The FX studio created over 700 shots for the futuristic anthology series. The art direction of the series was inspired by the paintings of Swedish artist Simon Stålenhag.

The post Tales From The Loop: VFX-Breakdown von Rodeo FX first appeared on DIGITAL PRODUCTION and was written by Patrick Poti.

When giant, vicious alien monsters emerge from a crevice at the bottom of the sea, humanity is faced with the question: resistance or extinction? They decide to fight the “Kaiju” (Japanese for giant monsters) with robots called “Jaegers”, which are brain-controlled by two pilots. This works quite well for a few years until ever larger monsters emerge from the sea and only four combat robots from America, Australia, China and Russia are left. In order to avert the apocalyptic inferno after all, the four nations enter the final battle for the world.

Industrial Light & Magic (ILM) was responsible for around 1,600 VFX shots, including “Jaeger” robots and “Kaiju” monsters. It was also the studio’s first collaboration with Mexican director Guillermo del Toro. Filming took place at Pinewood Studios in Toronto, where the film team occupied all the halls with the real film sets – the set of the main cockpit of a robot, the so-called “Conn-Pod”, was also located there (you can also find out more about the design and set construction of the Conn-Pod in this issue of DP in the interview with set designer William Cheng).

The cockpit set weighed around 20 tonnes and could be moved and even shaken by a gigantic hydraulic tilting device, the “Mighty Mo”. The actors wore heavy boots that were permanently installed on the platform. For the actors, the shoot was like a rollercoaster ride, during which they were sprayed with around 5,000 litres of water per take.

The actors’ ordeal was a great reference aid for the VFX artists, as the ILM team was able to transfer the actions of the actors in the robot’s head to the movement of the 3D robot models. The team deliberately decided against motion capture for the animation, as the “Jaegers” were not supposed to move too humanly or organically, but rather robotically. The main protagonist of the robot squad, which is explicitly described as female in the film, is the 80-metre-tall “Gipsy Danger”. Around 20 model versions of her alone had to be created for the different degrees of destruction. The secret stars of the film are the monster creatures with bright blue maws, the concept of which involved some of the best monster designers in the world. Both ILM and compositing artist Michael Ralla describe “Pacific Rim” as the most challenging and exciting project of their careers to date. We spoke to Michael Ralla about the reasons for this statement and the extensive pipeline and workflow changes ILM made for the project.

DP: How many VFX artists were employed by ILM and how many studios in total?

Michael Ralla: The work was split between the headquarters in San Francisco and ILM’s offices in Singapore and Vancouver as well as some smaller outsource vendors like BaseFX in Beijing, Ghost in Denmark and RodeoFX as well as Hybride in Montreal. In total, the team involved was around 650 people strong, and they worked on a total of 1,566 shots.

DP: How did ILM organise the workflow for the project internally?

Michael Ralla: Efficiency was extremely important. The overarching principle was: “Every dollar goes straight to the screen.” In San Francisco, ILM processed 432 shots, which was sometimes the most complicated part of the work.

As with previous projects, the teams were organised into departments. But this time they worked much more closely together – direct communication was actively encouraged and required. Whenever a certain combination of different artists seemed to work well together, the production tried to keep this constellation together: For example, as compositor I worked together with VFX-TD Florian Witzel and LightingTD Michael Rich on a whole series of similar shots.

The supervisor for the entire project was John Knoll, with co-supervisors Lindy DeQuattro and Eddie Pasquarello working together on the supplied work – but all three were based in San Francisco. The lynchpins of each day were dailies and nightlies, with the Vancouver department seamlessly involved via video conferencing. Basically, ILM in San Francisco shared a pipeline with Vancouver, which is mirrored in Singapore. From a compositing perspective, it was interesting that ILM had leads for each discipline and character. This was incredibly helpful as the “Jaeger” robot and “Kaiju” monster looks were really complex. Especially in terms of specific glows, materials and lights, which of course had to look consistent throughout.

DP: Which scenes were you responsible for?

Michael Ralla: I worked mostly on two specific sequences: the dramatic “Dockfight” between “Leatherback” and “Gipsy” and the exciting final fight between “Otachi” and “Gipsy” in the streets of Hong Kong – both incredibly rewarding tasks with shots that were a lot of fun to shoot.

DP: How was deep compositing used for this project and what are the benefits?

Michael Ralla: Deep compositing was used extensively, but selectively on a per-shot basis and only for things where it really made sense. As already mentioned, efficiency was very important because the storage space requirements and the computing power needed for a deep workflow are enormous.

With all the atmospheric effects such as water running on the objects, ocean and wave interactions as well as the consistent rain in almost every shot, deep compositing helped to avoid time-consuming re-renders due to holdouts that no longer fit, for example in the event of an animation change.

We also used the entire standard deep toolkit in Nuke, for example, to see where the renders were located in 3D space and punch them out correctly: In almost every shot, tons of Practical Elements like sparks, dust, smoke, explosions and of course rain were integrated, which helps tremendously with realism. In order to place these efficiently and correctly, it helps a lot to see exactly where the 3D deep elements are located in 3D space.

DP: What tools did you work with on “Pacific Rim”?

Michael Ralla: All the compositing was done exclusively in Nuke with the addition of a number of regular commercially available plug-ins and, as always, some proprietary ILM tools. There was no need as a compositor to write any pipeline scripts. All I did in that framework were small helper scripts that automated or facilitated repetitive things in my own Nuke scripts.

DP: How was Katana used in this project?

Michael Ralla: Katana was used almost exclusively by the lighting TDs, who also did quick slap comps of their passes in Nuke for testing purposes.

DP: Why did ILM primarily switch to the Arnold renderer for “Pacific Rim”? For which elements did you have to supplement the pipeline with other renderers?

Michael Ralla: I’m not a lighting TD, nor did I have anything to do with the pipeline planning. So unfortunately I can’t say much about this, except that the primary renderer for all characters was Arnold. V-Ray was used for the vast majority of environments, while RenderMan was used for water and simulation work.

DP: The final fight in the film between “Gipsy” and “Otachi” is a complicated scene: how did you manage the incredible number of render layers, especially all the reflections of the neon lights including the rain?

Michael Ralla: There’s a pretty simple answer to that: you just try to get the most talented people in the VFX industry to work on it (laughs). Absolutely right, there was an incredible amount of elements that mostly all interacted with each other in some way and were only put together later in compositing. The layering process was relatively complex, but all the people on the team did their best to facilitate this – and it worked pretty well.

DP: You’ve worked on many Hollywood projects. Why was “Pacific Rim” the most challenging and exciting project of your career to date?

Michael Ralla: The exciting thing about this project from a compositor’s point of view was the complexity of the shots and all the technical challenges involved, as well as some new approaches and ideas. Particularly in the compositing department, we used a whole range of new workflows, all of which were extensive but also highly efficient. It was really exciting to be at the forefront of this crew. At least in San Francisco, there wasn’t a single “easy” shot. As already mentioned, there were an enormous number of render passes, and dealing with the daily updates alone required good self-management skills.

DP: What was particularly challenging for you?

Michael Ralla: For example, I worked on the development for the final battle in Hong Kong: “Otachi” swims into a part of the harbour, causes a massive wave in the ocean, knocks over entire ships and finally smashes a car park full of exploding cars with the first steps ashore. So 15 seconds packed with dynamic action in a full CG environment where everything reacts with everything else – this shot kept me on my toes for quite a while.

But as I mentioned before, we tackled the challenge with a proven team of artists, and again it was Florian Witzel who took care of the fluid dynamics. There are simply synergy effects from this way of working, which makes things easier. The combination of two people in a team usually results in more than just the pure addition of the two talents.

DP: “Pacific Rim” is often compared to Michael Bay’s “Transformers”. What were the main differences between the two projects for you as an artist?

Michael Ralla: There are fundamental differences between the two projects, at least from a compositor’s point of view – not to mention the completely different mindset of the two directors, although they also have certain things in common, which I will come back to later. The most obvious similarity between the two films is the robots – but that’s about it. And even those were already largely different: 80 metre tall manned machines versus 15 metre tall intelligent robots. The overall style is completely different, it’s all about highly polished sports cars versus worn-out steam-punk machines and perfectly timed “golden hour” lighting versus constant rain and thunderstorms at night.

DP: And the compositing in particular?

Michael Ralla: “Transformers 3” had a much higher proportion of live action with shot material, because Michael Bay tries to do as much as possible in-camera, such as explosions and smoke. Realistically integrating the robots into the plate in this scenario is a challenging task, whereas in “Pacific Rim” the majority of the VFX shots were completely computer-generated.

The task there therefore centred on creating an overall image that looked photorealistic and plausible while still fulfilling all of the director’s artistic, stylistic and, above all, story-related requirements.

An extremely important task in “Pacific Rim”, for example, was to get the scale across in almost every shot. “Kaijus” and “Jaegers” are gigantic in size, but this is not always necessarily obvious to the viewer, which is why an attempt was always made to put them in relation to something of a known size. This is a little easier with the “Transformers”, which are derived from cars of a known size.

DP: What was it like working with VFX supervisor John Knoll? Michael Ralla: Working for and with John Knoll was an incredibly rewarding experience. John runs a pretty tight ship and knows exactly what he wants – which harmonised well with what Guillermo wanted. Knoll had a palpable feel for del Toro’s vision. At the same time, as an artist you still had quite a lot of room for artistic freedom. John trusts his artists and if an idea or suggestion makes a shot better and is realisable, John almost always agreed.

As overall supervisor of the whole project, he kept the crew happy by simply fully respecting and trusting each person. And despite the size of the team, he usually knew each person by name and was always approachable.

DP: Was del Toro actively involved in the post-production process?

Michael Ralla: Guillermo del Toro was definitely very involved and active in the post-production phase. He’s relatively similar to Michael Bay in that respect – they both direct the effects work as well as every other step of the film and there was no client supervisor in between on either “Pacific Rim” or “Transformers”.

Each week, Guillermo was shown both final and work-in-progress shots from John Knoll for approval – usually the artists would get direct and immediate feedback from him via video conference. It was definitely fun to work on his project and for him. He is inspiring and open to new ideas, but can still formulate precisely and decisively what he wants.

DP: How much freedom of choice did the VFX team have in designing the look of the digital elements?

Michael Ralla: Basically, working on “Pacific Rim” was more like working with the director than for the director: from the perspective of an artist doing shot work, we were definitely all in the same boat. Guillermo was an integral part of the team, visiting the crew often and talking to them directly and personally – he just wanted to get people excited about his film, his passion was infectious and definitely a big factor in the positive experience as a whole.

Del Toro has a colourful personality and the colour palettes for his sequences were just as colourful. “Find your inner Mexican” was a phrase he mentioned often – something I personally had to get used to. Especially as a German who, at least from an American perspective, stereotypically likes perfect right angles and a controlled colour spectrum …

DP: What else do you like about working at ILM that keeps you coming back for projects?

Michael Ralla: ILM, one of the global leaders in VFX, is definitely the most Oscar-winning company with probably the longest history in this field. Just getting a job interview there was a huge thing for me and then to be able to work there multiple times is still incredible. ILM values and respects its employees enormously, but the demands and requirements are high. For us Europeans, there are also relatively restrictive immigration conditions – but once you’re there, you’re rewarded with a great atmosphere, a collaborative and inspiring working environment, cool projects and incredibly nice colleagues.

DP: What are your next projects?

Michael Ralla: I’m currently finalising the “Ender’s Game” project at Digital Domain in Los Angeles, unfortunately I can’t talk about follow-up projects yet. However, looking at the current state of the feature VFX industry, I’m seriously considering – at least for a while – trying out the world of commercials.

The post Of “Jaegers” and monsters first appeared on DIGITAL PRODUCTION and was written by Mirja Fürst.