GPOY (thank God for a three-day weekend)

Some FX Notes (you know, just to remind people that this is an educational place): The exploding head scene was accomplished by filling a latex head with dog food and rabbit livers, and shooting it from behind with a 12-gauge shotgun.

mappeal:

Reaching for the Stars in the Real Effects of The Tree of Life

Senior VFX supervisor Dan Glass on working with Terrence Malick :

“The way Terry describes things is not visual, but emotional and often alludes to philosophy or music.”

At no point was the film to feel like a science documentary “or something that was presented, per say; He wanted it to be an emotional experience, to give the viewer a sense of wonder and beauty, the chance, the elegance of all these evolving processes.

But at the core he wanted us to pay due respect to science at the deepest possible level.”

via CGSociety.

The complex set piece was done in two four-day shoots – a “pre-crash,” where the kids can be seen making their own film, and a “post-crash,” after [production designer Martin] Whist re-dressed the set. Production captured many shots, both during and after the crash; for the actual impact, a total of nine cameras were used: four manned, and the remainder unmanned crash cams, such as Eyemos, placed in harm’s way (though all survived undamaged).

The sequence was planned out with Visual Effects Producer Chantal Feghali and Industrial Light Magic Visual Effects Supervisor Kim Libreri (with effects produced under direction of ILM effects legend Dennis Muren). ILM Animation Supervisor Paul Kavanagh created a simple animatic previsualization.

“It was mainly to block out basic action beats,” Libreri explains. “But J.J. had this great idea, given that it was such a large scale environment. Instead of pre-determining everything, he knew what the basic beats were, which he had drawn on little mini-boards. That was our beat sheet, to ensure we were shooting everything in the right order.”

Whist also built a 6-foot by 3-foot model, which enabled the team to envision where things such as cranes, crash cams and explosion events would be set.

“That was where we discussed the best angles for the camera, and where the kids could run,” Libreri recalls. Abrams still added cameras/moves on the actual day, insisting he didn’t want to bleed the scene of its reality through previz and storyboards. And Muren concurs, noting that a lot more was discovered on the day than VFX had anticipated. “You really want to leave directors and cameramen open to what they feel on the set,” Muren says. “Because the movie’s better that way.”

The crash, filmed on the “post-crash”-dressed set, involved capturing the “locomotive” (a green screen vehicle with a headlight, like that of the CG vehicle that would replace it) acting as a ram and smashing through the depot set. The ram was pulled through at a fairly fast clip (about 40 mph) by a cable attached to a crane.

Stationary cameras, of course, were not a part of the equation. In fact, Carr-Forster, shooting from the 50-foot Technocrane on the Maverick, was able to keep the scene’s focus on the kids, despite all the wild mayhem.

“So many other directors would use the explosion as the primary object,” Anderson says. “But J.J. is the kind of storyteller who uses the people as the primary object, and the explosion is almost secondary to the scene.” Anderson’s B-camera was on Steadicam on a dolly track, following both the train ram and the kids, and being pulled by grips moving at the ram’s speed. Libreri says it was like “a Ben Hur chariot Colin was on, to get some high speed motion, following the kids.”

(via Mystery Train : ICG Magazine / Showcasing the members of the International Cinematographers Guild)

Man, is it still a “miniature” if it’s over 40 feet tall? Insane-o. And they blew it up twice. (via)

fx guide: Can you talk about the style of animation? 

Dale Newton (sequence supervisor): We tried a few things to get a gritty and hand-made feel. The inspiration was the way Lotte Reiniger had the childlike-ness for her animation, but we didn’t necessarily want it to look like a stop-motion piece of animation. Because of the camera moves, we couldn’t always split things down on twos. So we couldn’t rely on any sort of technological roughing of the edges. The one thing we were keenly aware of at all times was the silhouettes which enabled us to play on a certain theatricality. The hands do so much of the talking - Death’s hands for example are almost as expressive as his face is. 

When we designed the characters, we tried to purposely design puppets. So we didn’t give the characters, for instance, whites in their eyes. You didn’t want to read them through, say, traditional blend shapes on their heads. It all had to be told throughout their entire pose. We were very conscious that the attitude of the characters at that point had to be read through their entire pose. This forced us into thinking theatrically, and we should feel like we’re watching a very clever puppet show, not a traditional character-animated movie. (via FX Guide)

[Please note: I saw a post somewhere about this on Tumblr before and tried looking for it to reblog but to no avail. I mean no disrespect to the original poster by posting this again, but if anyone knows where the original post is, please send it to me, and I’ll reblog the original and delete mine. Thanks.]

via:

What just happened? And Where’s Ron?

When time turns backward, Harry and his friends are filmed at 24 fps in the foreground, while in the background, four hours of hospital action is compressed into 10 seconds of time-lapse footage run in reverse. “Once we got the foreground, we were able to choreograph all of the background action” says [VFX supervisor Roger] Guyett. Adds [VFX supervisor Tim] Burke, “We had to work out a lot of mathematics to figure out how long it would take.”

“Basically, it was done as an 8-fps effect,” says Guyett, who worked on this with second-unit director of photography Peter Hannan, BSC. “We did some motion-control movement so we could control the time it took, but essentially, we just did a whole bunch of separate actions that we could break down and turn into 10 seconds on the screen.”

Lighting was the key to integrating foreground and background. During the time reversal, lighting changes had to proceed in lockstep. For the actors and hospital set, “we had lights outside the big casement windows set for daylight, moonlight and sunlight, all on dimmers and going through diffusion,” says [DP Michael] Seresin. These were mostly 10Ks and 20Ks, plus a “moon light” - a 20K HMI on a remote-controlled crane arm. “We also had one electrician handholding a light and walking onto a piece of wood which was on a roller on a pivot, so he could go higher. Very 19th century!

“As time starts going back,” Seresin continues, “we did the lighting changes for real. Guys on dimmer boards with split-second timing go through moonlight, dusk, magic hour, late-afternoon sunlight andmidday sun.” He adds with a laugh, “It happens so quickly it’ll probably only be the cinematographers of the world who’ll pick up on it.”

This progression guided the second unit’s background lighting. “But obviously, here the light would have to go backward,” notes Burke. “It got very complicated. In fact, it was probably one of the most complicated shots any of us had been involved in.” (via American Cinematographer)

inky:

Simulated wireframe effect from the film Escape from New York (1981).

What appears on those screens was not computer generated. Carpenter wanted hi-tech computer graphics which were very expensive at the time, even for such a simple animation. To get the animation he wanted, the effects crew filmed the miniature model set of New York City they used for other scenes under black light with reflective tape placed along every edge of the model buildings. Only the tape shows up and appears to be a 3D wireframe animation.

 Actor Joseph Gordon-Levitt Talks Training and Filming in Inception’s Various Revolving Rigs

What training did you do for the sequence? 

It was two weeks full-time, divided between basic physical training, combat, wire work and revolving corridor. On one hand it was incredibly challenging. On the other hand it felt like playing in my backyard as a kid. 

How did you keep your bearings? 

The corridor rotates in a constant rhythm, and you have to stay on beat. You have to know when the floor becomes the wall and the wall becomes the ceiling. I would count and repeat melodies—usually Bach—in my head to keep me on time. 

Were any parts more difficult than others? 

There are two rotating sets. The narrow corridor was more physically demanding, because you had to jump from floor to wall more often. But the larger room had less tolerance for error, because if you got behind the rotation, you could be looking at a 25-foot drop.