By Martin McEachern
The year 2007 would have been heaven on earth for Alfred Hitchcock. The great director—who once claimed to have shot every frame of his films before the camera rolled, and even hired artist Saul Bass to sketch every frame of Psycho’s shower sequence—would have been a kid in a candy store with today’s previz technology. Given the complexity of modern filmmaking, in which directors are forced to juggle live action, motion capture, and effects animation in a single shot, it is little wonder that previz is joining storyboarding as a routine station in the filmmaking pipeline. In fact, the development of all this past summer’s major blockbusters, from Spider-Man 3 and Live Free or Die Hard to Pirates of the Caribbean: At World’s End and Rise of the Silver Surfer, began with previz artists.
This group of growing niche artists worked closely with the directors on a conceptual level, to figure out the blockings and timings of the major storytelling beats; and also on a technical level, to help foresee the limitations of the sets, determine the right lenses or the length of a dolly track, or find the best placement of greenscreens, lighting, and camera equipment for obtaining the best coverage.
And if you think previz is strictly a preproduction step, think again. Previz artists are now working on location, helping the director cope with sudden and unforeseen changes with a live-action shoot, designing shots sometimes hours—even minutes—before the cameras roll, hence the term “durviz.” And when directors need a fast preview of how an unfinished character animation or set extension will composite into a live-action plate, they can turn to a “postviz” artist.
As previz, durviz, and postviz open up new frontiers in the industry, they are providing a creative outlet for a class of artist whose talents may not lie in creating the most realistic subsurface scattering shaders or the most perfectly sculpted blendshapes, but rather in storytelling and cinematography. A previz artist has to be an uber generalist, according to Brian Pohl, a producer at Persistence of Vision, a leading previz vendor. The previz process, he says, exposes the artist to every aspect of the CG pipeline, allowing the person to sidestep into a specialist career in modeling, animation, or texturing. Moreover, previz is a “hybrid” career, wherein the artist always has one foot in CG and another in live action. In fact, these artists interface so closely with the director and on-set crew that previz serves as the perfect training ground and starting point for making one’s own movies.
For insight into the development of previz and the creative and technical challenges facing these artists, we turned to three worldwide authorities on previsualization: Christopher Batty, lead artist at Pixel Liberation Front; Ron Frankel, president of Proof-inc; and Pohl, also COO and senior previz supervisor at Persistence of Vision. While all those houses utilize industry-standard packages such as Autodesk’s Maya and Softimage’s XSI, dedicated previz software tools are also beginning to emerge, among them Antics 3D and ILM’s forthcoming Zviz.
At the seminal point in the filmmaking pipeline, previz technology has the potential to influence how the production process unfolds downstream. And no one in the industry has expressed a greater interest in changing the way the industry works than George Lucas. After feeling “hamstrung” by a makeshift Maya-based previz system developed for Star Wars Episode III, Lucas charged ILM with developing the Rolls Royce of previsualization tools. The system would integrate seamlessly into the facility’s Zeno pipeline so that the studio could take its film from preproduction to postproduction on an ILM backbone. The Zviz system is currently being road-tested on the new Star Wars TV series, as well as on several feature animation and game projects.
“We designed Zviz to be available to the public, so it’s friendly and approachable,” says ILM’s head of R&D, Steve Sullivan. “George [Lucas] wants to change the way the industry works and creates movies, and he sees Zviz as an educational tool—one that will be accessible for people who are just telling stories as individuals. For it to have that kind of impact, we would need to get it out there in some nonproprietary form.”
ILM intends to sell a public version of Zviz but has not set a release date. In fact, the tool continues to evolve and iterate based who is using it at a given time.
Presently, Zviz has three modes: one for scene creation, animation, and shot setup; a viewing mode for reviewing the shot catalog; and an editing mode for assembling the sequences. While the software offers a selection of modeling tools and houses a library of animatable cards and generic props (such as buildings, cars, and characters), it also allows users to import full scenes from Maya, XSI, or other standard 3D modeling packages. “Zviz is not a full-on modeling tool. We want enough functionality to enable the user to get the idea across and set up the action,” says Sullivan.
To simplify things, user-friendly context menus make the software more accessible to those who are less-technically inclined. In addition, shots can be mocked up using a game controller, joystick, or other similar device.
According to Sullivan, one of George Lucas’s mandates was that users be able to quickly make a one-off correction to a master action without affecting the master action. While Zviz has digital analogs for industry-standard cranes, dollies, and other equipment, Sullivan says most users are only interested in working with a free-roaming camera, aspect ratio, and a set of lenses, thus leaving the selection of instrument rigs up to production.
For directors who need greater realism in their previz’d character animation, Zviz offers tools for creating blendshapes, sculpt deformers, and IK controllers, as well as supports physics and dynamics simulations (useful when laying out a room, for instance). Nevertheless, the software tries not to expose the authoring controls, assuming the user would, for example, create a set of facial shapes in Maya, XSI, or ILM’s Zeno, and drag them into the software.
“Zviz sits on top of the Zeno framework, where we do a lot of this work on movies like Pirates and Transformers. Therefore, the Zviz interface itself wouldn’t have all the sliders and controls for shapes, but the way it’s currently packaged, you can get behind the scenes and do whatever ILM can do,” says Sullivan. “Somebody using just Zviz would tend to import prebuilt rigs; they can create poses and facial animation via imported shapes and IK, and drag in precanned animation, motion capture, and keyframe as much as they want.”
Zviz and Mocap
While Zviz does not have a dedicated compositing mode, artists can accomplish this task via image planes and camera projections. Bluescreen extraction can be done with underlying Zeno operators. Where Zviz really shines, however, is on the mocap stage, allowing previz artists to composite live mocap characters into their previz’d scene in real time. “Zviz is the live front end that we use on our motion-capture stage (the scene and sequence recording tool for the mocap sessions); so you could create a scene and a set of actions in Zviz on your laptop, and then go down to the mocap stage and have live mocap characters running around in that environment, with your camera running live in that environment, as well,” explains Sullivan. “You can cut your shots and intermingle them with shots done offline. It’s completely seamless.”
Right now, Zviz relies on pieces of the LucasArts game engine to power its real-time lighting and rendering. “For example, it uses the way the materials are represented to give us a look beyond the normal OpenGL stuff,” says Sullivan, who, along with Proof-inc’s Ron Frankel, concedes that hyper-realistic, real-time lighting is still beyond the scope of current technology. Regardless, previz artists will still rely on key lights, backlights, and fill lights, as well as intensity and direction, to visualize certain dramatic moments.
While ILM intends to expand Zviz’s physics simulation and real-time lighting and material capabilities, Sullivan cautions that most clients find such minutia in previz distracting. At one end of the work-style spectrum would be Steven Spielberg, who prefers rough animatics that loosely depict the major beats of a scene. At the other end would be David Fincher, whose Panic Room was so heavily previz’d that camera movements were planned to the millimeter and recorded in a shot bible that was religiously followed on set. In retrospect, Proof-inc’s Frankel, who did Panic Room’s previz, concedes that such rigorous specificity may have been overkill, not to mention creatively stifling for the on-set crew.
ILM does not draw any distinction between previz, durviz, and postviz, and has designed Zviz to function at any stage in the Zeno pipeline. “We see Zviz as not just a preproduction tool, but as the backbone of a shot that you can duck in and out of at a given time,” adds Sullivan. “Our matchmoving system is in Zeno, along with our image-based modeling stuff, compositing operators, and a lot of the pieces that you would need for a shot, so a previz scene can exist at any point in the pipeline. You could take a water simulation or a particle field that was generated by ILM’s high-end simulation tools and have it playing back within your previz scene, even though you didn’t create it in Zviz. So, for us, that kind of distinction between pre-, dur-, and postviz doesn’t apply.”
Nevertheless, ILM has designed Zviz’s scenes to be fully portable to and from other packages, so that if a studio chooses the tool for previz but not as the final effects vendor, the asset sharing is seamless. Obviously, there would be some loss due to translation or missing features between the software.
Make no mistake, Zviz is becoming the root of George Lucas’s new production paradigm for the medium, influencing the way shots are conceived and created on set. “That was George’s intention; he wanted to work differently and needed a tool that would allow him to do so,” says Sullivan. “It supports all the standard ideas of previz.”
For a glimpse into the typical workday (and work night) of a previz team assigned to a major motion picture, we turned to Proof-inc, which helped lay the groundwork in many of the big action sequences for this past summer’s Live Free or Die Hard. “We did previz for over four months during preproduction, and then we overlapped for a number of weeks into production itself,” says Ron Frankel, president of Proof-inc.
One of the most challenging sequences to previsualize takes place in a freeway tunnel, wherein hero John McClane launches a car into a hovering helicopter. “We worked closely with the second unit director, the special effects team, and the stunt coordinators to map out a number of cool car-crash action beats,” says Frankel. “We previz’d McClane tumbling out of the car, another (featured in the trailer) of McClane and his young sidekick, Fowler, standing in the middle of traffic as the car comes hurtling through the air toward them.”
Using the previz as a guide, the on-set crew executed most of the shots practically by launching real cars on a massive catapult constructed on set. Even the helicopter is real, prior to the CG takeover for the explosions, shattering glass, and debris.
“The director’s goal was to rely on physical effects as much as possible. So after we designed the sequences, we’d analyze them to determine how far we launched the car, how many feet it traveled through the air, and then take that information to special effects to see what they could do,” explains Frankel. “Then, they’d come back and say, ‘We can only launch a car 60 feet through the air,’ so we’d go back and re-engineer the shots within the limitations of what the stunt guys thought they could do on the day of the shoot. There is so much back and forth, hand-in-hand work with second unit and special effects.”
Working in Softimage XSI, Proof-inc’s artists built models of the locations based on surveys. As they blocked the shots, the production designer and art directors would study the previz shots closely to determine what the camera was and wasn’t seeing. The group also worked closely with the art directors and set dressers early on to be sure that enough scenery was visible to sell the location to the audience.
While on set for the tunnel-crash sequence, Proof-inc was also tasked with doing real-time compositing to help compose shots hours before filming. “The director obviously couldn’t have the actors in the shot when they were launching the cars. Therefore, the crew placed a camera in a crash box on the road and bluescreened the actors in,” recounts Frankel. “After the director and crew chose a take of the car, we tracked it, did a little motion-control move, and exported that move to our previz. Then, we composited our previz actors on top of the plate and showed that to the director and the visual effects supervisor to get them to sign off on the size and placement of the actors.” As Frankel points out, all that work was done on location with limited prep time.
While Proof-inc previz’d the sequence using generic cars and street props (the crew didn’t know which props would be available during the shoot), their modelers became more accurate for the climactic Imperial Highway sequence that pitted McClane’s big rig against the villain’s F22-B plane. “We got incredibly accurate with the proportions and dimensions of both those vehicles. We were working in a confined location, so their sizes were critical to the sequence design,” says Frankel. “But the limited space is what makes the sequence so exciting.”
For both the tunnel crash and the Imperial Highway sequences, the team used 360- or 180-degree panoramic plates shot on set, as well as stock footage to flesh out the backgrounds. Photographs of the distant backgrounds were mapped into the previz to provide an urban context and a sense of scale without having to clutter the scene with heavy geometry.
Both action sequences entailed high-impact damage effects for shattering glass, folding metal, scattering debris, and crumbling concrete, and in the early previz stages, the team experimented with rigid-body dynamics to get a feeling of windows smashing. “We are finding we’re using rigid-body dynamics more frequently in previz,” says Frankel. “At the same time, unless you’re really willing to invest the time in perfecting the simulation, it’s easier and faster to hand-animate it. Because of the speed with which a production unfolds and beats and shots change, you could invest a lot of time in a rigid simulation that worked for one shot but would be useless if, for instance, a car were suddenly replaced with a truck in a crash sequence.”
As with most previz houses, Proof-inc avoids proprietary software as much as possible to be in sync with the VFX house. In this instance, the previz team completed all its animation in XSI and the motion and camera tracking in 2d3’s Boujou, enabling the group to move the camera and characters around to get the right composition.
Furthermore, a previz house must learn to adapt quickly to each director’s previz style, which is as unique as the person’s approach to directing. For example, when Frankel worked at Pixel Liberation Front (PLF), his crew created a detailed blueprint for every shot of David Fincher’s Panic Room. During the postmortem, the group discovered that a lot of it was either ignored or followed too literally.
Just as a previz house must adjust to a director’s previz style, previz artists must be equally adept at adopting the final vendor’s software package, workflow, and methodologies at the outset of production. Frankel provides a case in point: “For Die Hard, we did all our work in XSI before the studio finally awarded the project to The Orphanage, which works in Maya, so we had to translate all our scenes.” On the other hand, with Pirates of the Caribbean: At World’s End, ILM was chosen as the final vendor before the studio hired Proof-inc for previz. ILM provided its character rigs, so Proof-inc’s animations would fit correctly into the pipeline. ILM even assigned one of its character animators to the previz team for a number of weeks, showing the best way to animate the rigs so they would be as compatible as possible with the studio’s pipeline.
Mocap for Previz
Proof-inc’s artists are currently hard at work on a martial arts film they cannot name, for which they’re exploring the largely untapped potential of using motion to re-create the choreography from videotaped footage so the director could block out his camera moves. “Instead of cutting up the action with insert shots, the director had all these great ideas for using dynamic, radical, continuous camera moves that would capture the energy of a martial arts fight,” says Frankel. Instead of painstakingly rotoscoping the videotaped actors, Frankel persuaded the director to do the rehearsals on a motion-capture stage at Giant Studios in Playa Vista, California. In a single day, the director, actors, and previz team did 10 or so fight sequences, and Proof-inc imported that data into its previz scenes within XSI.
Once the director had perfected his shots, he had the stunt performers re-enact the fight on location, matching the camera moves on set to the ones worked out in previz. “Using mocap in previz is something I’ve always been skeptical of because, unlike camera data and actor blocking and staging, it can’t be passed down through the production pipeline,” says Frankel. “What’s the point of mocapping an actor if they’re just going to walk on set anyway? It is pointless if it’s not going to save you money downstream. But, this project really opened my eyes to its potential.”
Similarly, PLF also has been quick to embrace the potential of mocap in previz, resulting in a new type of work paradigm. Until recently, PLF’s experience with using mocap in previz consisted of importing mocap data provided by forward-thinking productions. The best example of this, says PLF executive producer Sean Cushing, would be the freeway sequence from The Matrix Reloaded: “We were given mocap of a choreographed fight sequence, which we then composited into the larger world (on top of a moving truck), figuring out how best to shoot and cut the action with the rest of the sequence.”
That being said, most productions are not prepared to mocap large sections of action in the early stages of production, when previz is so very important. In fact, they’re often not willing to lock down choreography that early on, and rightly so, since the sequences are still in their infancy. Previz is usually in the position of having to work out the choreography of the action and the camera in accordance with the wishes of the director and the various department heads responsible for the sequence.
Also, previz sequences are more and more looked to as tools to work out emotional content and overall storytelling. Therefore, the performance of the previz characters has had to improve dramatically over the years. However, compelling character animation takes time, a luxury not afforded previz, which has to be fast and iterative. So, the quality and emotional impact of detailed character animation is typically at odds with the previz process and the needs of the production schedule.
To address those issues, PLF has integrated motion-capture technology into its pipeline (see “PLF Makes a Big Move in Previz,” pg. 29). The company even owns a Vicon mocap system that it has recently used to generate previz animation for Jon Favreau’s upcoming Iron Man and numerous other projects. Additionally, PLF has assigned one artist to the production of the upcoming The Incredible Hulk film as a dedicated mocap previz artist. “We’ve been utilizing previz mocap data on nearly all of our current productions,” says Cushing.
As part of the process, PLF also captures secondary characters and alternate motions as much as possible to allow for fast changes in a sequence without having to acquire new motion. Compared to traditional previz character animation done on the same timescale, the quality of animation from the mocap system is orders of magnitude more refined. However, when such realistic motion is on the screen, it is quickly accepted and taken for granted, allowing the filmmakers and previz artists to focus on the filmmaking rather than on the animation. Setting Up Scenes
While Pixel Liberation Front tackled postviz for Live Free or Die Hard, the facility collaborated with director Sam Raimi during all three phases of production for Spider-Man 3 (see “Facing the Darkness,” May 2007, pg. 8). Despite the efforts, even the best laid previz planning for the film was often upended by the harsh realities of shooting. For example, Raimi and PLF had planned to shoot the crane disaster sequence at a certain skyscraper, but on the day of the shoot, the crew could only run the cable cam along the opposite side of the building from what had been planned in previz. So, at the last minute, PLF designed the shot so it could be flopped backwards, scaling the whole world to be flipped along the axis line.
For these cable-cam shots and crane moves, PLF delivered to the camera department detailed diagrams (resembling architectural drawings) showing the camera’s path and the placement of equipment, such as cranes, tracks, and the cables, which were modeled, animated, and simulated. PLF lead artist Chris Blatty, along with two other artists, worked on set with Sony Pictures Imageworks’ John Schmidt, who was responsible for the technical planning of the cable-cam shots. Because both used the same software, with the same tools and assets, changes could be made quickly.
In the postviz phase, PLF also met continually with Raimi to check shots and plates, make placeholder composites with temporary effects, and allow him to experiment. PLF’s temporary composites and effects not only showed the final effects vendors what Raimi was looking for, but were incorporated into the film for early test screenings.
Combining Spielberg’s loose approach to previz and Fincher’s preference for specificity, Raimi works in both extremes during previz. On very complex shots, he was able to use previz to devise a rigid plan for a shot. “Other times, he was free enough to make up a shot on the spot,” says Batty. “To a certain extent, previz can let a filmmaker be spontaneous. If you’ve already spent a lot of time designing something upfront, then that allows you to stray, because you already know what you need for the sequence.”Asset Sharing
Spider-Man 3 is also a perfect example of seamless asset sharing between the previz house, PLF, and the final effects vendor, Sony Pictures Imageworks. “We were able to use all their assets, including characters, rigs, and digital sets, albeit at a slightly lower resolution. Since we were using the same software (Maya), we could simply hand over our scene files for them to use directly,” says Blatty. The previz team was also creatively intertwined with the art department, helping to guide the set design prior to and during production. If Raimi requested changes during the shoot, PLF could make those changes in previz and hand the scenes over to the art department to see exactly what the director was asking for.
The quality of character animation expected in previz is also increasing exponentially, especially in animation-intensive films such as Spider-Man 3. While the increasing emphasis on facial and body performance is obviously intended to help capture the emotional aura of a sequence, Batty notes that there is an element of salesmanship to previz that pushes an artist to go beyond what’s necessary. “For instance, music and sound effects can also give the previz a bit more emotion,” he adds. “Movies are, after all, about people, not things.”
A common sentiment among all previz houses is that the advancement of previz is currently encumbered by two technological barriers: the lack of real-time lighting and rendering power and the failure of cameras to operate in real time like their real-world counterparts. For this latter reason, the cinematographer is still somewhat alienated from previz. “Real-time lighting is still not good enough to specify a lighting or color scheme through previz,” says Batty. “For rendering, we continue to use some methods from gaming. We tend to rendermap most of the sets and hero props. This method bakes in a lot of the shading/lighting and makes it easier for the OpenGL captures that we do. We hardly ever use software renders for the shots, since this can increase render times from seconds to hours.” POV
Although the term “postviz” has been coined only recently in previz circles, no company has become more acquainted with its challenges than Persistence of Vision (POV), which recently completed 800 postviz shots for the upcoming Journey 3D, a remake of Jules Verne’s sci-fi classic Journey to the Center of the Earth. “Postviz artists tend to be a little more technically minded than their previz counterparts. If you can’t track shots, your effectiveness as a postviz artist is severely limited,” explains senior previz supervisor Brian Pohl.
Journey 3D comprised numerous sequences of actors directly interacting with digital characters. Director Eric Brevig would develop his shots with a storyboard artist first, then pass those boards onto POV’s artists, who would construct the digital assets and environments and discuss the need for compliance between the previz sets and the real sets. If such compliance was unnecessary, POV would “blue sky” the environment and focus on the storytelling element of the previz sequence, a concept Brevig terms “conceptual” previz. He refers to the matching up of the digital environments with the physical stage and set as “technical” previz. With the former, you are selling an idea to a studio; with the latter, you are executing the idea, he adds.
Journey 3D is shot in stereoscopic 3D, and part of POV’s previz work was to help Brevig explore the stereo potential of various sets. To do so, POV customized Maya to produce left- and right-eye versions of the shots, thereby achieving complete control over interocular distance and convergence. Aside from the custom stereo camera created for Journey 3D, POV has developed other custom tools to streamline its previz process, including masking and slating tools in Maya.
Pohl estimates that 98 percent of POV’s work is done in Maya using polygons, which he says is a better medium than NURBS. “When we model assets for previz, such as the underground environments in Journey 3D, the level of quality is similar to game cinematics,” he notes. “The model should be built to scale and be a fair representation of the character or object used in the production.” POV was activated six weeks into the preproduction of Journey 3D, allowing the artists access to the production designer for previewing and re-creating the set designs prior to production.
Sometimes, however, blueprints, photographs, and models from the final effects company are not available for postviz set re-creation. Then, the company pushes forward by tracking the shot in Boujou and camera-mapping plates onto rough geometry, as was the case for Evan Almighty. In one scene, POV had to re-create a courtroom set entirely from existing blueprints; no digital sets were available.
Artists fully rigged Journey 3D’s previz characters in Maya, equipping them with basic facial expressions and phonemes, as well as specialized props. Blendshapes are also commonly used at POV, as are weight maps, particle and dynamic systems, and Maya’s Paint Effects. POV’s artists occasionally used hard-surface dynamics to simulate various crashes and collisions during the subterranean journey.
“With Maya’s new dynamics engine, Nucleus, the process is even easier,” says Pohl. “There are few technologies we don’t use in previz; it’s just a matter of implementation and whether or not it will ‘buy’ us what we need during short notice. The only thing we steer clear of is a software renderer or technologies that can’t be seen or exploited in OpenGL.” As a result, nearly 95 percent of POV’s previz rendering is done in hardware, though Pohl is quick to point out that improvements in real-time lighting and rendering are surely needed in this area.
While Journey 3D’s previz work was largely conceptual, there were a few instances when it was used to provide technical solutions for the DP on set. “We had to solve stereographic questions and calculate specific speed and motion for the camera,” says Pohl. POV has several rigged camera systems that mimic the motions and limitations of several common rigs. Typically, the facility also has the DP’s lens package available when it starts the process.
The challenge of providing durviz services on set, maintains Pohl, is having a system in place that is fast enough to composite previz characters and set elements into a shot right on the fly. To that end, POV has been working with some motion-capture technology companies, and Pohl sees other technological developments on the horizon that will broaden the power of his tool set. “Enhanced OpenGL cards are always improving, and new motion-tracking tools are emerging as well. Auto-rigging tools, Maya’s Paint Effects, and FBX have been considerably helpful, too,” he adds. “Essentially we have to create an entire micro pipeline in previz. We must model, rig, and animate nearly everything, so I always look for tools that make that process faster. Previz is all about speed.”
While POV often shares its previz files with the final vendors, Pohl is quick to warn against trying to achieve such seamless asset sharing between the two houses, which could effectively turn previz into a layout department. “They are two different animals. Layout trades speed for accuracy, and the more pipeline-dependent you get into a VFX company’s system, the more likely you’ll never be able to achieve your true previz objective, which usually is conceptualizing shots,” he says. “However, it’s always beneficial to receive files from the VFX house. Using low-rez proxies of their actual files is a big plus, and if that data is available, we try to take advantage of it. But, beware of diminishing returns: Sometimes files coming from the various VFX houses implement scripts and plug-ins that a facility doesn’t have access to and the VFX company will not grant access to.”
Despite the usefulness of geometry caching and the platform-independent FBX file format, POV and other previz houses stress the importance of adopting the software of the final vendors, which could undermine the efficacy of dedicated previz packages such as Antics 3D and even ILM’s Zviz (which POV is beta-testing) on high-end projects.
While Maya, Softimage, and Zviz tend to dominate in the budgetary stratosphere, smaller productions, first-time directors, and even a few Hollywood heavyweights are turning to other, less-expensive previz packages to visualize their films and get them funded or green-lit. These include RealViz’s StoryViz, Frameforge 3D’s Studio, Autodesk’s MotionBuilder, and Antics Technologies’ Antics 3D.
Built around 3D game technology, Antics 3D enables filmmakers to build complex 3D sets, animate actors and cameras, and edit shots. While Antics now supports the importation of objects from Autodesk’s 3ds Max, most users draw upon a massive library of props and actors (which come equipped with their own textures and animations) to furnish their sets. Rodney Charters, DP for the hit TV show 24, currently uses Antics 3D to block his shots prior to going on location, while Dean Devlin, writer-producer of Independence Day, calls it the backbone of his preproduction process. With a price of $595, Antics 3D is a fraction of the cost of a high-end package and poses only a fraction of the learning curve.
Antics 3D has three modes: set creation, rendering, and animation/audio/camera choreography. Actors can be directed point-and-click style (without the need for keyframing), or animated with motion-capture data, either imported or drawn from a library. At present, the lighting system supports multidirectional lighting with all variations of the color spectrum. Spot lighting is not yet supported.
“We see Antics as a general visualization tool, not just a specialist previz application, whether for film or broadcast markets, or for the legal, security, and education fields, and even the huge corporate and consumer markets,” says Fred Medina, Antics’ VP of sales and business development.
For aspiring filmmakers and animators driven by their passion for storytelling, directing, and cinematography, previz may seem like a dream job. Asked to describe what they looked for when evaluating a previz artist, each company echoed the need for a strong generalist. “We like our artists to be capable of modeling, animating, character rigging, character animation, and camera animation. They should be capable of understanding film language because they must interact directly with the director and the DP. A solid understanding of cinematography is critical,” says POV’s Pohl. “I also prefer artists who are fast and efficient at what they do, but also possess a certain degree of attention to detail. We have to send files to a number of VFX houses, and if their project files are a mess, it tends to annoy the facilities.”
PLF’s Batty also looks for someone with a good eye for shot-making and camera work. “Previz artists are a part of a small commando team that is responsible for anything in a sequence,” he says. “You don’t know what problems you’ll be tackling next. It could be a simple close-up or a complicated shot with a lot of particle effects and dynamics. The only thing you do know is that you’re there to make some cool shots that help tell the story.”
Proof-inc’s Frankel concurs: “We look for someone who is going to have a good sense of story and composition. It’s not enough to be a good animator; so much of filmmaking is about composition and how to move the camera around in order to tell a story, so we’ll even look at people with a background in cinematography, painting, or photography.”
Previz is not a stepping-stone to becoming an animator; animators focus on other things, particularly if their emphasis is character work. Rather, previz artists focus on cinematography. If anything previz orients the artist in two directions: One is technical, into the field of layout and technical direction, and the other is creative, into more of a filmmaking discipline or directing. To this end, a number of schools are offering previz training, including The Art Institute of California, Los Angeles, Otis College of Art and Design, and Gnomon School of Visual Effects, where POV’s Pohl has taught and attended classes; he has also hired artists from the school.
“Orient your demo reels to the type of work you want to be hired for, which might require separate reels. Presentation is key, so spend a little money to make sure your resume and demo reel look good and perform as expected. Don’t leave people guessing,” advises Pohl. “Previz is not an entry-level job. You are expected to be both a computer artist and a fully capable cinematographer as well, so study film.”
Such advice would make Hitchcock proud, knowing that the group of artists he helped spawn would be returning to his work decades later for inspiration.
Martin McEachern is an award-winning writer and contributing editor for Computer Graphics World. He can be reached at firstname.lastname@example.org