How to Train Your Dragon is made marvelous by that rarest of creatures – a nuanced and relatable CGI animal. Part dragon, part puppy, Toothless can convey a range of emotions with a curled lip or a twitch of the eyes, resorting only occasionally to the safety net of anthropomorphism. It’s the corniest of stories – a wimpy kid shows that brains trump brawn, and that gruff warrior types do not hold a monopoly on courage and persistence. The strongest message is that enemies are not always what they seem, and might be prey to the same fears and pressures as you are. It’s exactly what you want your kids to believe, but not necessarily what you want to see cynically mobilised to flog a Happy Meal. Dragon also benefits from the most effective use to date of 3D technology. Foreground and background are really unhooked from one another, dragons seem suspended in the air before you, and the Viking village depicted becomes a bustling perspectival pile-up of objects, high cliffs and big faces. I was also reminded of the close friendship between 3D movies and flying sequences. A film about dragons, and learning how to ride them, naturally lends itself to scenes of hurtling through the sky at breakneck speed, a white-knuckled passenger on a flight of vertiginy. See, for example, how Robert Zemeckis souped-up Dickens’s A Christmas Carol by having Jim Carrey’s Scrooge thrown through the air at regular intervals:
[See also the follow-up post, Digesting Avatar.]
Do you feel like the game has changed? Are we in a new age of spectacular cinema, freed from technological limits? That’s what was promised, but has Avatar rescued us from our humdrum lives of everyday movies with everyday special effects? My initial verdict is, well … sort of.
First things first (and here’s where you’ll find the greatest concentration of potential poster-quotes) Avatar looks astonishing. Really. It has wondrous moments when you momentarily accept the tangibility of the lanky blue folk on the screen, and it makes perfect sense that these are couched in a narrative about a man exploring a new world via a new body: Cameron meshes together the diegetic events and the experience of their spectacles perfectly, so the spectator’s exploratory view of Pandora (where the film takes place) can be focused on discoveries of plants and species that are, at the same time, discoveries of CGI novelties. It means you don’t have to feel bad about stopping and staring: it makes gawping at stuff feel like a plot point. But the plot is so stale that it might even be seen as a deliberate strategy to choke off any sense of suspense or complexity and force the audience to focus on the immediate splendour of the present moment: don’t worry about what’s going to happen, just check how good it looks as it’s happening.
Another Robert Zemeckis performance-capture movie, and that can only mean one thing – another round of reviews complaining that the “actors”‘ eyes all look a bit funny. You know the angle by now (if not, see my earlier post about Beowulf): Zemeckis, director of Back to the Future and Who Framed Roger Rabbit, makes a film using digital avatars of popular movie stars (Tom Hanks in The Polar Express, Ray Winstone, Angelina Jolie and Anthony Hopkins in Beowulf), and critics scratch their heads over why they bothered spending all that money if the CG versions are going to look just like the actors themselves, and bemoan the lack of spark and vitality in the eyes and faces of the digital figures. It’s all part of the mainstream discussion of digital technology, with a widely held distaste for the unnatural, uncanny, inhuman or just plain cold sensations evoked by the experience of watching these souped-up game sprites
First of all, let me say that, though I was skeptical about 3D, I’m officially sold on the idea. I don’t really need 3D, and it doesn’t add much except giggly novelty to a film like A Christmas Carol, but if it’s there I’m going to enjoy the fad while it lasts. There’s an argument that Zemeckis’ new film shows Victorian London off splendidly with its 3D fly-by shots that spatialise the setting and give a vivid sense of period. The title sequence is a swooping, one-take marvel that shifts its attention from the macro of a snowy city’s rooftops and the micro of rosy-cheeked carol singers and mischievous children. It’s spectacular, but it’s such a technical showcase that it feels incredibly modern, canceling out the evocation of a past time or of any of London’s historical specificity. It’s a bit like an animatronic display in a museum – it might give you a fun summary of a time and place, but it won’t transport you, and it’s no substitute for reading a book on the subject. That said, the shots of 3D snow are a wonderfully simple effect, and it’s tempting to try and catch a snowflake on your tongue; that must be the ultimate validation of a special effect – it had me leaning forward with my mouth open like an imbecile (there’s a poster-quote in there somewhere, I’m sure…).
Zemeckis’s animators have finally solved the problems of unnervingly uncanny digital actors, at least as far as Jim Carrey’s Scrooge is concerned. Click on the image above to get a hit of the full detail of the graphics. You’ll also notice that on close inspection you can see the joins between separately programmed elements like the cloth hat and Scrooge’s forehead: digital animation makes hypercritical geeks of us all! But the face is extremely enabled in its facial movements. The CG Colin Firth is, by way of contrast, a waxy monstrosity that fails to mime the jauntiness of the actor’s voice – that may be one of the problems of hiring stars for this job: you picture the face that you know goes with the voice, and the digital version can only be found wanting. Other supporting parts, and a lot of the extras, seem to have received much less care and attention than the star, and they still have the dead-eyed roboticism of earlier mo-cap efforts. Things are getting better though – a close-up of Bob Cratchit’s wife fighting back the tears is pretty close to being affecting (high praise, eh!), and the CG turkey looked really delicious. Scrooge is caricatured just enough that he is recognisably human enough, but twisted and stretched enough to seem effectively cartoony. There are almost no jokes in this film, leaving Carrey to wring humour out of the physical exertions of embodying tight-wound stinginess and all-consuming self-interest, sucking everything into his tight frame as if conserving all of his energy for himself. His spine is so arched it’s like he’s slowly bending back on himself.
But performance capture was not really about making it easier to control facial expressions – that’s just a problem they need to iron out – instead, it’s part of Zemeckis’ totalising system of drag-and-drop world-building that allows him to craft completely malleable locations and performers. It means he can wring a series of value-added performances out of Carrey (in the process implying that the ghosts, all played by Carrey, are creations of his tormented psyche), making him into a poseable action figure with enough freedom that they can still trade on his name.
Perhaps the biggest surprise about A Christmas Carol is how reverent an adaptation of Dickens’ story it is. Sure, there’s a bit more flying through the sky in this version, but most of the dialogue is Dickens’ own, and the designs of the ghosts are drawn from John Leech’s illustrations from the 1843 edition:
A Christmas Carol is such a familiar story that I wish they had taken more liberties with the source material and reconfigured it like they did with Beowulf; because the animation comes to life in the big action set-pieces, it gets leaden during the dramatic scenes when the actors are asked to perform tasks devised for humans – it doesn’t play to the strengths of an all-powerful hunk of CGI. When Scrooge is dragged into the grave, it’s a fabulous bit of baroque bombast, but when he has his change of heart, there’s just not a lot of joy leaping off the screen. Maybe that’s the answer – mo-cap works fine for actors, as long as they stay grumpy.
[See also Toy Story 3: All Things Must Pass]
The most startling thing about the new 3D version of Toy Story is that it seems perfectly designed for the format. Objects poke out at the camera and there are point-of-view shots: check out the scene where Woody and Buzz are carried into Sid’s house, peering out through a gap in zip of the bag they’re trapped in. The teeth of the zip frame a deep view into the house, while Sid’s dog Scud snarls and jumps at the camera. The tension between foreground and deep space seems to have been designed for 3D, as does the thrilling climactic chase, but no shots have been changed for the re-release. And because Pixar didn’t go for any gimmicky in-your-face 3D tactics in the first place, the story is still allowed to breathe without the distractions of peekaboo spectacle. Some of Randy Newman’s songs (“Strange Things” and “I Will go Sailing No More”) are still distractingly literal in describing (or dictating?) the onscreen montages (though all will be forgiven by the time he gets round to the emotional whallop of “When She Loved Me” in Toy Story 2 – I’m welling up just thinking about it…), and the Svankmajer-for-kids grotesqueries of Sid’s room are still an imaginative highpoint even for Pixar. The animation, groundbreaking though it was fifteen years ago, has dated slightly, particularly in the animation of human figures, and most tellingly in the character of Scud: they simply didn’t have procedural modelling programs in place to generate realistic fur, so poor Scud looks a little plastic. (Take a look at this feature about Pixar’s RenderMan system for a glimpse of how much the company innovated in the development of animation.)
Some things haven’t changed at all, though, and Toy Story still impresses by the sheer elegance of its plot: every scene blends imperceptibly into the next step in its development without revealing the mechanics of how it will manouevre all of its characters along the formulaic steps of its life-lesson journeys. The film flies by not just thanks to its breezy, witty script (and peerless voice cast), but because there’s not a moment of slack or digression from its simple tale succinctly told.
So, what’s with the 3D and how on Earth did they do it? You can hear director John Lasseter talking about it here, and read more about the conversion to 3D at the New York Times. Of course, its not an automatic process, even when dealing with digital elements (which are surely easier to convert than live action footage); a second camera has to be placed in the virtual space of the film to create the illusion of depth, and a team of “stereographers” had to select which parts of the frame to pull out and which to push back, and how far to push or pull them (by varying the distance between the two “cameras”. Often, that must be a fairly logical choice, bringing foreground elements out and pushing the backgrounds into the distance, but it’s interesting to hear how lead stereographer Bob Whitehill made some of those choices:
When I would look at the films as a whole, I would search for story reasons to use 3-D in different ways. In Toy Story, for instance, when the toys were alone in their world, I wanted it to feel consistent to a safer world. And when they went out to the human world, that’s when I really blew out the 3-D to make it feel dangerous and deep and overwhelming.
Thankfully, the effect is not distracting, and never used to excess: Pixar have not cheapened their work with a gimmicky makeover, and judging by hushed Saturday matinee crowd with whom I watched it, Toy Story still has the ability to enthrall. Being armed with the foreknowledge that its sequel will be even better doesn’t hurt, either…
Apparently, we’re in the midst of a 3D revolution. What do you mean you hadn’t noticed? It’s mostly confined to cartoons and kids stuff, being a gimmicky plaything: at the far end of my cynicism scale, 3D is the most expensive game of peek-a-boo in history. I’ll sit this one out until the novelty wears off and we can see if it’ll have any staying power in the long run. Many attempts have been made to revive it as a crowd-pulling tactic, often to refresh a stale franchise or genre, sometimes to reassert the primacy of a big-screen cinema experience in response to competition from television or digital piracy. Louis Lumière even showed a stereoscopic remake of L’Arrivee d’un Train to the French Academy of Science in 1935 showcasing the “new technology” (stereoscopy itself predates cinema) by referring back to the first moments when films amazed audience with a purported illusion of depth and presence. It might be interesting if 3D could get over the hurdle of novelty and become just another part of the filmic furniture (I have a lazy eye and wear glasses, so the effect is sometimes wasted on me anyway, so I won’t mind if it disappears altogether again); 3D films so far have mostly been “fun” or spectacular. Who will dare to use it for a harrowing drama or discomfiting, miserabilist horror?
I didn’t come here to bury or to praise 3D. Actually, I came with the humble intention of a quick post about Jackie Chan’s 1978 vehicle Magnificent Bodyguards. Jackie Chan starred in five films in 1978, one of which was Drunken Master, a film which would show the successful formula of mixing martial arts and physical comedy and end forever the attempts to turn him into the new Bruce Lee, but you can tell that this early effort, dates from a time when his handlers were shopping around for a consistent persona for their gifted young star. I hadn’t previously been aware that it had been shot in 3D, but the evidence was unmistakable. Objects are thrust towards the camera during combat scenes, suddenly popping into sharp focus a couple of inches from the camera. If 3D is supposed to be used to spatialise an environment, Magnificent Bodyguards uses it for jolting effect. The close-up shots are never held for more than a second, making for a rhythmically impactful use of the technology, as if you’re supposed to forget about the 3D until it pokes you in the retina. Of course, I was watching this on a 2D DVD, and inferring most of these effects from the ways in which the apparent 3D shots are incorporated into the action.
As objects lunge out towards you, usually in a shot where the combatant looks directly into the camera, they come into privileged focus – they take on a singular objectness for a fleeting instant. It might be overused to the point where that specialness is depleted, though. The problem with 3D is that it makes literal or hyperbolic some of the effects which action films are trying to achieve anyway – they want you to feel closely involved in the action, physically engaged as you flinch in your seat. That usually has to be achieved with expert performance and rhymthic editing. Look at these two sequential shots from the film, in which Jackie Chan ends one shot by punching towards the camera, before the reverse shot shows the punchee falling away from the camera:
The shot/reverse shot technique normally suggests a spectatorial space around which the action can occur: you might seem to be sat around a table with a group of people, without actuallg being the table, if you get what I mean. In this case, the “gap” between the two shots is compressed so that the spectator’s position is crunched between them. Are you expected to become both the head being punched and the fist punching it? Maybe the martial arts film is one of the few genres where 3D makes sense, since it aims to reduce the distance between viewer and drama, to make you react bodily to the performers on screen (even if it’s to note the contrast between their activity and your passivity), so it’s only a matter of time before somebody gives it a try in a non-Panda-based film.