In 2012, unsuspecting movie audiences were introduced to Ted, the animated teddy bear star of Seth MacFarlane’s irreverent comedy of the same name. Ted was a hit at the box office, rapidly gaining status as one of the biggest-grossing R-rated comedies of all time.
Now, the potty-mouthed plush toy is back in the sequel, Ted 2. The new film sees Mark Wahlberg reprising his role as Ted’s buddy and original owner, John Bennett, and chronicles the bawdy bear’s attempts to prove that he has the same legal rights as a regular human being … while amusing and offending pretty much everyone he encounters along the way.
Key to the success of both films was the convincing creation of Ted himself. For the sequel, as for the original movie, animation and visual effects duties were divided between Tippett Studio and Iloura, with additional support from Weta Digital, and with Tippett Studio’s Blair Clark fulfilling the role of production VFX supervisor.
Tippett Studio delivered around 600 shots of Ted, including a break-in sequence, a scuba dive, and a parody of John Candy’s singing performance with Ray Charles in Planes, Trains and Automobiles. Iloura delivered approximately 1,000 animation shots during their nine-month assignment.
For this exclusive roundtable Q&A session, Cinefex brings together insights from the following key artists at Iloura and Tippett Studio:
- Eric Leven, Visual Effects Supervisor, Tippett Studio
- Glenn Melenhorst, Visual Effects Supervisor, Iloura
- Colin Epstein, Senior Compositor, Tippett Studio
- Brian Mendenhall, Animation Supervisor, Tippett Studio
- Jeff Price, VFX Editor, Tippett Studio
- Howard Campbell, Lead Technical Director, Tippett Studio
- Niketa Roman, PR Specialist, Tippett Studio
Now, without further ado, let’s talk teddy bears!
On the original Ted, the visual effects load was shared between Iloura and Tippett Studio. Is that how it worked on the sequel, and was the original team an automatic choice for the production?
GLENN MELENHORST: The old team were definitely brought back together for Ted 2. On Ted, we shared Ted around fifty-fifty with Tippett. This time around, Iloura completed twice the number of shots as we did for the first film, sharing the rest with Tippett and a small portion with Weta.
ERIC LEVEN: I think there’s always a hope that you’ll be asked back, but you can never assume it’ll be so – ask Boss Films about Ghostbusters 2! There’s always love and friendship in Hollywood, but there’s always money, too. So we had to bid for the work and meet their price. Certainly Tippett and Iloura had a leg-up, but it was never a given.
GLENN MELENHORST: We were approached from the start to be part of Ted 2, and it was down to our availability and the bid. Seth was so happy with the work on the first film it’s no surprise he wanted us and Tippett back on this next adventure. As with the first film, Blair Clark was show VFX supervisor. He is an awesome supervisor from Tippett and he really understands both Seth’s aesthetic and how we work here at Iloura. As a company with a strong creature animation focus, we have always felt culturally aligned to Tippett and really enjoy working with them.
How closely involved was Seth MacFarlane with the visual effects process?
ERIC LEVEN: I think it’s fair to say that Seth IS Ted in many respects, so he was instrumental in Ted’s performance. He not only provided motion capture, voice acting, and video reference, but also detailed animation performance notes. In addition, he always made sure that Ted stayed on-model.
GLENN MELENHORST: Seth was closely involved in reviewing our dailies from start to finish. Given his background in 2D animation, his focus was more on the subtleties of the animation than the nuts-and-bolts FX side of things, while Blair made sure the work between the studios matched and all the technical aspects of the job were on track.
Jumping back briefly to the first film, how was the original Ted character designed and rigged to give him that authentic teddy bear look?
GLENN MELENHORST: The original Ted was based on some rough designs from Seth. When it came to making the actual model, we collected several teddy bears and studied them to make sure our asset had all the right seams and panels. We had to make sure the whole way from model to rigging to grooming that we were creating a plush toy and not an animal or cartoon character.
ERIC LEVEN: The model and rig are generally pretty simple. The face rig is a bit more complex to make sure that we can get the correct range of expressions from Ted. What makes his face challenging is that he can’t move his eyes, his nose doesn’t change shape, and he only uses his eyelids for a very occasional blink. We rely much more on the eyebrows and mouth but we – and Seth – are constantly watching to make sure his face isn’t too misshapen or otherwise off-model.
GLENN MELENHORST: We also put a lot of irregularity and asymmetry into him so he would feel like a toy who’d been around for thirty-odd years and been beaten up a bit.
Did you develop a new Ted for the sequel, or did you just bring the old bear out of storage?
GLENN MELENHORST: I guess from a production perspective, and an audience expectation standpoint, Ted needed to be his old lovable self. But, as you can imagine, every film teaches you things about animation and pipeline, and after the first film we were keen to reinvent some of our workflows and rendering tech to make shots easier to turn out, as well as to step up the look a notch.
ERIC LEVEN: We were re-treading old ground in the sense that this is a different story about the same characters. Mark Wahlberg is back, and Ted is back. We don’t want a different Mark Wahlberg, and we don’t want a different Ted.
GLENN MELENHORST: Another thing to keep in mind is that software and hardware continue to march forward, speeding things up and allowing us to discard some of the cheats and workarounds and use a more unified raytrace lighting pipeline. In terms of animation, the process was much the same, although this time we went into the show already understanding much of the nuance that Seth was after for the character of Ted.
Specifically, what changes did you make to the digital teddy bear for Ted 2?
GLENN MELENHORST: We made a few nips and tucks, mainly just tidying a couple of things up which were bugging us in the first film, such as messy hair around the lips which interfered with lip-sync. We reworked his facial rig slightly to make him a little easier to animate, and also his hands were reworked slightly to allow him to articulate a bit more easily.
ERIC LEVEN: The real changes we made at Tippett Studio involved upgrading our shading model to allow for more realistic fur lighting. Because of this, we didn’t need to futz with Ted too much in the comp, and he looked basically correct right out of the render. On Ted, there was a lot more playing with the look in the comp, which was obviously not ideal. Tippett Studio also transitioned to Katana on Ted 2 – that provided a great deal more flexibility for the TDs, and much faster turnaround of lighting tests and changes.
HOWARD CAMPBELL: We used environmental images captured on-set to mimic subtle variations in light and colour. This really improved our ability to integrate Ted into the scene and tie him in with the actors. We saw real leaps in details such as fur quality and reflections, but were still able to maintain the look and feel of the bear from the previous film. It’s the same bear, only better.
GLENN MELENHORST: The biggest changes we made at Iloura were improvements in our rendering technology. In Ted, we used a hybrid pipeline of raytracing in V-Ray and REYES in 3Delight. This time around, we still opted for a hybrid approach but put more emphasis on 3Delight, which we now also used for raytracing. We also completely reworked our cloth pipeline, using Marvelous Designer to build as well as simulate our clothing. This gave a very robust and realistic result, even when Ted was doing outrageous actions such as in the main title dance routine.
Watch the featurette Ted 2: A Look Inside:
Animated characters have benefited hugely from improved flesh and muscle simulations. Is that of any use when you’re animating a stuffed toy?
ERIC LEVEN: For a teddy bear, we didn’t need any muscle sims. What brought the model to life was the use of cloth simulations to get the right shape and weight of his body beneath the fur.
GLENN MELENHORST: We did do a full cloth solve to simulate the effect of his body being made of fabric. This meant if Ted bent over or twisted his body, you would get creases and folds appearing.
Were there any special requirements for Ted’s fur?
GLENN MELENHORST: On the original Ted, we had a scene with Ted in the bath. That required custom groom and shaders. This time, there were no real special requirements for his fur … other than to look real!
ERIC LEVEN: The effects model was basically the same on this film. There were new costumes to deal with – a raincoat, a scuba suit, a hooker outfit – that provided their own aesthetic challenges, but nothing technically new.
Stuffed toys often pop up in horror films – there’s something inherently creepy about seeing them come to life! How did you get around this problem?
GLENN MELENHORST: I think dolls in horror films are either motionless or spin their heads slowly with creepy music, which makes them seem more threatening. Ted moves more or less like a human, and he talks and cracks jokes, which is enough to give him empathy.
ERIC LEVEN: Our very first test of Ted for the first movie DID look creepy! We wanted Ted to appear worn and ragged – the way you might imagine a thirty year-old stuffed animal would become. So we had a few rips that had been sewn up and a big ratty cloth patch that covered a hole. His fur was matted and his expression was just generally pissed off, even when he was joking. Looking back, all these things made Ted look pretty creepy. The key was to make him look like a more ordinary stuffed animal that wasn’t quite so raggedy. Worn and old, yes, but not dirty. We also played his expression more sardonic and less angry. That helped us relate to him more instead of being repulsed.
What about Ted’s eyes? They’re really just blanks. How do you stop them looking dead?
COLIN EPSTEIN: It would be fair to say that Ted’s eyes got more attention per shot than just about any other element – except perhaps his fur. Since we’re so trained to focus on someone’s eyes to glean emotion, we did everything we could to make his eyes not just look physically real, but to really add to the intent of the scene. Every choice was based on that goal, whether he needed an angry glare, a stoned vacancy, or a mischievous twinkle.
GLENN MELENHORST: Naturally, having no whites in the eyes means Ted has to turn his whole head to make eye-lines work. With no moving parts to his eyes, we had to achieve sidling glances, stares, astonishment, and so on, with body language and subtle eyebrow animation. It’s very much like Snoopy or other characters who have dots for eyes: the tilt of a brow can really communicate a lot. With a character like Ted, you can have him just stare, completely motionless, and if you add welling music and a slow push in, the audience will conclude that he’s sad. You animate your intention, and the audience fills in the blanks.
How important is the way the eyes reflect their surroundings?
COLIN EPSTEIN: We had reference from a “stuffy” for almost every shot, but we only used that as a starting point. Each sequence environment got its own reflection set-up for Ted’s eyes, using set data and images, and then we’d manipulate that based on the feel his eyes needed. We’d often remove reflected details that were really there but proved visually distracting. For instance, the Comic-Con sequence was predominantly lit with a grid of lights on the stage ceiling. But accurately reflecting those rows of bright pinpoints in Ted’s eyes fought with his eyeline and cluttered things up visually, so we played those down. A key look note throughout the show was to make sure Ted’s eyes looked like old plastic, instead of brand new and shiny, or biologically alive. The practical bear had very clean and shiny plastic eyes, so we used them mostly for reflection and kick position reference.
How did you achieve the “old plastic” look in the renders?
COLIN EPSTEIN: Ted’s eyes were heavily treated in the comp passes. Since we knew going in that adjusting Ted’s eyes was always going to need specific attention, the comp scripts had a standard set-up that let the compositors control just about every aspect of each eye separately. A stock row of blurs, colour grades and transforms was there in Nuke when a shot was started. If the kicks landed smack in the centre of Ted’s eyes because of the angle of his head, it often gave him a creepy look we called “devil eyes”. In those cases, we would push the kicks off-centre in the comp to keep him appealing. Another issue was that Ted is modelled wall-eyed, so kicks and reflections in one eye would look drastically different in another. This sometimes made it hard for him to look like he was focused on something specific. So elements in his eyes would be pushed around a bit to counter that. Little adjustments like these happened throughout the show to strengthen Ted’s performance.
For the original Ted, Seth McFarlane performed scenes off-camera in a motion capture suit, with a stuffy used as a placeholder in shot. Was the same methodology used for Ted 2?
GLENN MELENHORST: Yes, it was very much the same this time around. Seth’s mocap setup – called Moven – provided us with reference for Seth’s upper body, principally his arms and the angle of his head. The system gave no facial feedback, or legs or body motion. We used the mocap as a basis for keyframing, adjusting the data to fit Ted’s body and modifying it where needed.
JEFF PRICE: For each camera angle, we would receive both a stuffy pass – with the VFX supervisor acting out the scene with the stuffed bear – as well as a greyball pass. We could then compare the stuffy size and placement in each shot with our animated Ted, as well as the light and shadows with the greyball in the scene.
GLENN MELENHORST: Whereas Seth was able to shoot a sideways glance, Ted needed to turn his whole head to look sideways. Often, therefore, the mocap was used only as reference. Combined with captured video, it gave the animators all they needed to understand Seth’s intention for the shot.
JEFF PRICE: As was the case with the first Ted movie, Seth McFarlane’s mocap performance generally served as a rough first animation pass, with our animators altering and fine-tuning Ted’s movements into a believable performance.
Animators are also actors. Is it important to “cast” the right animator for a particular shot?
GLENN MELENHORST: Animators are definitely actors. That’s something I believe quite strongly. We find a lot of animators who have skilled-up through online training courses, or through some of the bigger animation studios, tend to have learned what I can only describe as “Chaplin-esque 1920’s silent era acting” – overly expressive, hyper-extended, vaudevillian extremes. This has a place in many children’s films, but is not so useful when animating Ted. One of the best animation reels I ever saw was a study of a guy reading a newspaper. The subtlety was so well-observed – I loved it. Ted required that sort of observation, because his performance was so stripped back that every tiny key mattered.
BRIAN MENDENHALL: We spend a lot of time trying to keep consistency between performances. It is very important to the believability of the character as a whole. The worst note you could ever get from the director is to say that he wants to see the animator from one shot handle the animation on another (which did happen once). We don’t want the client to ever think there are any individuals behind the scene. I’d like to think you can’t spot the difference when you watch the movie. Maybe I can – but not you!
GLENN MELENHORST: More importantly than spotting differences between animators, we worked to not let the audience detect differences in Ted between studios. Matching renders and the quality of light and integration is one thing, but continuity in animation across vendors is another beast all together. Naturally, at Iloura we can tell a Tippett shot because we didn’t do it – but I hope it’s a harder task for the audience.
What was the most difficult scene you had to work on involving Ted?
GLENN MELENHORST: One of the greatest challenges was a shot where Ted fights a goose. The shot was turned over in the last two weeks of delivery. We had a simple goose asset provided, but we had to rig, animate and feather the goose in short order – no small feat. Also, it was a lengthy shot. We accomplished it by dividing the animation between animators and splicing the acting together. As well as it just being a difficult shot to turn around, Seth was still actively blocking out ideas on the shot during the last week. That meant we needed to run everything in parallel as much as we could, building our lighting pipeline on blocking animation, re-rendering as new animation was fed through the pipe, and compositing with whatever was at hand, just to get the thing done.
And the most complex?
GLENN MELENHORST: The title sequence of the film centres on a Busby Berkeley-style dance number with Ted out front of a hundred or so dancers. The main set piece – an oversized plywood wedding cake – needed clean-up and roto to remove screws and scuffmarks, so we spent time making it look pristine. The same was true of the dance floor, which was mirrored but scuffed, and while we didn’t clean it out entirely, we did remove a lot of the visual clutter while retaining the complex reflections of the dancers. This extended to painting out camera rigs and light rigs, and digitally extending the curtains and floor.
As well as cleaning up the environment, did you also have to edit any of the dancers’ performances?
GLENN MELENHORST: In the last shot of the sequence, the camera tracks back and up as the dancers form a perfect triangle. The girl on the furthest right didn’t hit her mark, and so the back right point of the triangle was a mess. One of our 2D guys, Johnathan Sumner, spent about a month reconstructing the shot by moving the girl, restoring all the arms, legs and flowing gowns that were originally across her, filling in the gaps left by her being slid sideways, and fixing the reflections.
Describe how you got Ted dancing in the opening number.
GLENN MELENHORST: Ted’s animation for the sequence had three real influences. Firstly, we matched mocap provided by the choreographer for the sections where they knew what they wanted. Secondly, there were times when we needed Ted to copy the dancers behind him and dance in sync. Thirdly, totally custom animation was needed when Ted was not the focus of the shot, when he needed to backflip or tumble or act in a way not possible to capture, or simply when the choreographer left it up to us to come up with the performance. We blocked the dance, refining and revising as the choreographer and Seth edited and revised the sequence.
One member of our animation team, David Ward, has a strong background in dance and really took to the sequence. It was a perfect fit, and he and the others on the sequence turned out some really great animation. It was not enough simply to match the dancers or use the choreographer’s mocap, because Ted is so small. We needed to exaggerate his silhouette, poses and timing, snapping up large gestures to help him really pop.
Our lighting team had the task of not only matching the studio lighting, but also creating believable reflections in the scuffed floor. They had to make Ted glow as a bright dress passed him, or darken him when he was shadowed by the passing dancers – we rotoscoped many of the dancers to isolate the jungle of legs. The studio lighting also changed colour and intensity through the shots, all of which all needed to be tracked and accounted for. The compositing team had the arduous task of interleaving Ted behind all the dancers, blending shadows and putting the final spit and polish on the shots. Overall it was a unique sequence and a great team effort.
We’ve heard a lot about Ted himself. Are there any memorable VFX shots that don’t feature the bear?
GLENN MELENHORST: The opening shot of the film is a cosmic zoom from the Universal logo orbiting the earth down into a stained glass window of a church in Boston. The shot was one of the first to be turned over and one of the last to be finished. As in the first film, we wanted the camera move to be a single unbroken shot, with no wipes or cheats between takes as is common in these things.
How was the cosmic zoom shot put together?
GLENN MELENHORST: Satellite and aerial photography was sourced and purchased, and our FX lead, Paul Buckley, began to block the camera move along with our match-move department. As the camera approached the city, we decided on what to build in 3D and what to leave as 2D matte painted elements. Eventually, we settled on building ten city blocks of Boston, along with some outlying skyscrapers. This build included all the vegetation, cars, pedestrians, kids on swings, debris in gutters, and so on. Many months of modelling, texturing animation and matte painting went into the shot, and we are pretty proud of how it turned out.
Thanks to everyone for contributing to this Q&A. Any closing thoughts on Ted 2?
GLENN MELENHORST: Well, I do want to say what a pleasure it was for us at Iloura to work with Seth and Blair again. Iloura and Tippett are like sister companies – we really enjoy working together. Seth is always so complimentary about our work, and it is amazing how much that enthuses everyone in every department to put in a huge effort. Likewise, Blair was always so generous with his praise and support, and we all count him as an honorary Ilourian!
NIKETA ROMAN: I know it’s been said, but we at Tippett Studio really want to emphasise what a strong director Seth MacFarlane is and how great it is to work with him. He’s got a background in animation and a very strong vision of who Ted is as a character. He speaks the language of an animator and communicates what he wants very clearly. That clarity and ease of collaboration is really what makes Ted so successful … it’s also just plain fun to work on a movie where you spend so much time laughing. We had a great time!
- Tippett Studio
- Weta Digital
- The Foundry – Katana
- The Foundry – Nuke
- Moven Motion Capture
- Chaos Group V-Ray
- Marvelous Designer
Special thanks to Niketa Roman, Fiona Chilton, Simon Rosenthal, Ineke Majoor and Anna Hildebrandt. “Ted 2” photographs copyright © 2015 by and courtesty of Universal Pictures, Media Rights Capital, Tippett Studio and Iloura.