On Coraline, in 2009, Laika printed about 20,000 replacement faces and then had to paint every one of them. The printer laid down a single colour of resin, so freckles were a brush job — reporting from the studio puts one of Coraline’s faces at over a week to finish. That cost set the ceiling. You can only have as many expressions as you can afford to paint.
Wildwood has more than 142,000.
The printer is only the output device
The faces come off Stratasys J750 and J850 PolyJet machines, which print in full colour by jetting photopolymer in layers. Getting a painted sculpt to survive that trip is the interesting part, and it is handled by software called Cuttlefish, built at Fraunhofer IGD. Cuttlefish takes a coloured 3D model and works out how to stack translucent resin so the thing that comes out matches the thing the artist approved on a monitor.
Anyone who has fought a texture through a render and then through a print knows exactly what that is. It is a colour pipeline. Resin is not ink and it is not a display, it scatters light under its own surface, and making a subsurface material hit a target albedo is a problem I have watched eat whole afternoons on jobs with none of this at stake.
So the “handmade” story has a large software stage in the middle of it. Each face starts as a digital sculpt, gets painted and rigged, and only then becomes an object. The animator picks up something physical. What they pick up was specified in a DCC package first.
142,000 of anything is a database problem
My sympathy goes out to whoever ran that department. The studio says it does not reuse expressions, so those are 142,000 distinct parts. Each one belongs to a character, a frame, a shot. Each one has to be findable on a shelf by a human being in a hurry, and traceable back to the sculpt it came from when a note comes in and the whole set has to be reprinted.
Nobody gives you an award for that. It is naming conventions, version control and a shelving system, and if it fails the animation stops. The published figure of 142,030 comes from a joint Laika and Fraunhofer release, so treat the precision as promotional — but the order of magnitude is not in doubt, and the order of magnitude is the point.
Around it, the old crafts held. Armatures hold the poses. Magnets hold the faces on. Eyes still get nudged by hand, frame by frame, because nothing has beaten a thumb and a dental tool for that yet. Over 2,700 shots. More than 220 puppets — some outlets say 230, and Laika has not published a number I would stand behind. Roughly 231 weeks of animation. Sixteen years from Colin Meloy and Carson Ellis’s book to a screen.
That pattern repeats whenever a production decides the camera should see a real object. HBO’s Lanterns built a real ring suit instead of leaving it to CG and spent the money on the build rather than the comp, for the same reason: the lighting is free and correct when the thing exists.
The number that decides whether Laika does this again
$12.2 million domestic, opening weekend, per IndieWire’s Monday analysis — third for the weekend. Earlier Sunday estimates had it nearer $11.6 million, so the figure firmed up as actuals came in. An A CinemaScore. PG-13, which for an animated feature is a bet in itself.
The comparisons cut both ways. That opening is roughly level with Kubo and the Two Strings, which is the house’s beloved film and not its profitable one. It towers over Missing Link, which opened to $5.9 million and limped to $17 million domestic. Fathom points out that Laika’s five previous features averaged about 3.6 times their opening weekend at home, which lands Wildwood somewhere in the low-to-mid $40 millions domestically.
Against a reported budget near $100 million.
International is still to come and a film like this has a long tail, so that is not a verdict. But hold those two numbers next to each other and you can see the shape of the problem for anyone whose job depends on this studio. The pipeline worked. The reviews are the best the studio has had. The arithmetic is still tight.
The film that beat it to the top of the same weekend was Clayface, which took about $24 million — and which this blog liked for a related reason, that the practical makeup carried the film where the digital melting did not. Two films, one weekend, both of them better in the shots where something physically existed in front of the lens. One of them cost a fraction of the other to make that true.
Sixteen years is a long time to spend proving that a printer can hold a colour.