Behind the Seen

by Charles Koppelman

  An Editor’s Tools

  What does an editor use to think about the film? Every editor has his or her own methodology. For Murch, it’s as if he builds himself a personal radio to tune in to the film using several different frequencies: picture boards with up to seven key images from each setup captured and mounted on black foamcore; two sets of script notes—one written on viewing dailies, a second on re-screening footage before editing a scene—which are embedded in his FileMaker log book; and handmade scene cards, coded by color and shape to delineate plot trajectory and the flow of characters.

  So now, on the third day of filming in Romania, with Sean Cullen getting the Final Cut Pro systems plugged in and working upstairs at Kodak Cinelabs, Murch begins preparing his scene cards.

  July 17, 2002, Murch’s Journal

  About 1/4 way through scene cards. I am doing them by hand this time, easier to look at. What color should Inman’s trek be?

  On his last few films Murch prepared the scene cards using FileMaker on the Macintosh. For Cold Mountain he goes back to making them by hand. “There is something appealing about the visual handcraftedness,” Murch says later. “The personality of the handwriting is more engaging to the eye, especially if I’m going to stare at them for a year and a half.”

  July 18, 2002, Murch’s Journal

  Working still on scene cards. Amazing how long it takes always. But good I am doing them by hand this time, not computerized printing.

  The scene cards, picture boards, and script notes are simple and uncomplicated tools. But they aren’t just different methods of cataloging. Like composer and printmaker John Cage’s throwing the I Ching to determine creative choices, these tools allow Murch to incorporate randomness into the edit process. If a scene isn’t working for some reason that isn’t readily apparent, a sideways glance at the picture boards might reveal a hiccup in the pattern of images that wasn’t obvious before. Let’s reshuffle the scene cards and see what color pattern emerges. Or, a simple script note (“alt line reading”) made in dailies may turn out to unlock a solution—as the alternate line reading, “He’d kill us if he had the chance,” did for The Conversation. But it requires forethought and effort to plan for the unplanned, to invite the unexpected, and to prepare these alternate tools for working on a film. Taking two days to prepare handmade scene cards at the start of production may seem extravagant. But film directors who embrace Murch’s style know that if they find him in his editing room cutting card stock into odd little shapes, what looks like child’s play is really serving the best interests of their film.

  * * *

  DigitalFilm Tree Gets Paid

  Back in Los Angeles, the same day Murch begins making scene cards, DigitalFilm Tree finally receives reimbursement for purchasing Murch’s Final Cut Pro systems. Ramy Katrib, in a digital home movie, turns the camera on himself and speaks with typical understated calm as he walks from the edit room to his office: “The Miramax wire transfer came in,” he says in a close-up so extreme it distorts his face. “And just in the nick of time—we had $532 in our account.”

  * * *

  Murch’s scene cards in progress for Cold Mountain.

  “Blue with a yellow background means Inman is in a scene; plain blue means Inman is not in that scene. A lot of blue cards in a row means not much Inman—which makes me wonder ‘is that a good idea?’ A triangle indicates I feel it is a pivot scene. The size of card equals the approximate length of a scene.”

  Ironically, the more techno-centric film editing gets, the more powerful Murch’s custom-made innovations become. The organic qualities of the scene cards and photo boards compensate for perspectives that are hidden in the digital world. The efficiency, speed, and increased choices of non-linear editing all have their benefits. But systems like Avid or Final Cut Pro obliterate some film editing tasks that contribute to the editor’s creative process. As Murch often points out, the simple act of having to rewind film on a flatbed editing machine gave him the chance to see footage in another context (high-speed, reverse) that could reveal a look, a gesture, or a completely forgotten shot. Likewise, the few moments he had to spend waiting for a reel to rewind injected a blank space into the process during which he could simply let his mind wander into subconscious areas. With random-access, computer-based editing, a mouse click instantly takes the editor right to a desired frame; there is no waiting, no downtime—and fewer happy accidents. The photo boards are one way to compensate for this.

  Murch assembles his picture boards as film dailies become available. These postcards from the film provide him another tool to see the movie.

  “With computer-based editing, there is no downtime—and fewer happy accidents.”

  Film-based editing is now referred to as “destructive editing” in contrast to the “non-destructive editing” of computers, also known as non-linear editing. When a scene is edited the old way—on film with taped splices—the only way to see if it might work better with, say, a different order of shots (the wide shot at the end of the scene instead of the beginning, for example) is to pull the scene apart at the splices and rejoin it in a different order. Not only does this take time to execute physically, the first version now no longer exists (unless a videotape copy was made along the way). In digital editing, the scene can be re-cut over and over again, and all the versions are preserved for later consideration. The computer program plays the footage by sending a request to the hard drives where the original media is stored: “play setup 293-take 4, from frame a to frame b.” With a couple of keystrokes that command can be changed to: “play take 293-take 5, from frame c to frame d.” Both versions still “exist” since no media has actually been touched (destroyed).

  Digital editing, being so versatile and fast, provides film directors with many more choices than they have in film-based editing. When a director comes to the digital editing room to look at a newly re-cut scene, he or she can see the new version, the old version and—after a few clicks, have the editor play a third, recombined version of the other two.

  Much of an editor’s energy, and most of the assistant editor’s, is spent organizing, logging, tracking, storing, and—like the sorcerer’s apprentice—trying to stay in control of a huge volume of material. To get an idea of how much raw film footage is at the editor’s disposal, recall the shooting ratio—the amount of film shot, compared to the amount of film used in the final theatrical release. It can range from 10:1 in low-budget independent films to 100:1 and more, depending on the style of the director. Stanley Kubrick, for example, was notorious for doing a hundred or more takes of the same shot. At the other extreme, Clint Eastwood is often satisfied after one or two takes. If a shooting ratio is 50:1, and the running time of the finished film is two hours, that means there will be 100 hours of raw footage, or 540,000 feet of film (16 frames of film equals one foot, or .66 seconds). That’s just over 100 miles of film!

  With so much footage to organize, it’s no wonder that in film-based editing chunks of footage often get lost. The missing pieces are usually “trims,” or snippets of film only a few frames in length. These little bits multiply as a film is winnowed down to its essential self, shedding microseconds here and there as the editor finds the truest 1/24 of a second for a shot to begin or end on. The black hole that swallows film is called the trim bin, a square metal container the size of a shopping cart, lined with muslin to minimize film scratches, where the trims are hung on hooks.

  A standard film trim bin.

  But trims fall off their hooks in the commotion of editing and disappear into the hamper like socks in the laundry. Having to search around the bottom of the bins for a missing trim was an assistant’s tortuous, painstaking job. One of Walter Murch’s contributions to the film editor’s tool kit, the Murch Bin, helps minimize this problem of footage vanishing from sight.

  All editors use an accounting method to keep track of film footage. There is no one, industry-standard way to do it; each editor devises a system that suits his or her n
eeds. Murch put his logbook together in FileMaker Pro, a database program that can search, find, calculate, and view by using multiple fields of information, customized to his preference. Each record represents one continuous camera take.

  The “Murch bin” uses the ends of crochet hooks instead of straight metal pins. Rubber bands make trims even more secure. Bicycle wheel clamps allow an editor to raise or lower the cross-beam to suit her needs.

  A page from Murch’s Cold Mountain scene list kept in a FileMaker Pro database.

  JULY 23, 2002, BUCHAREST, ROMANIA

  It is a week after principal photography began, and as film footage flies into the second floor at Kodak Bucharest, record keeping begins in earnest. Now that he has eight days’ worth of material, Walter makes his first estimation about how much film volume he can expect, based on a shooting schedule of 110 total days.

  July 23, 2002. Murch’s Journal

  354 Records as of today which represents 8 days of shooting. So 45 records per day x 110 days = 4955 or almost 5000 records at the present rate, first and second unit both using multiple cameras ¶ Tuesday, September 3, 2002 that rate is still holding: 45 a day, at 103 feet per take, so average 5000 feet of dailies a day. Grr. That would be 550,000 feet at the end of it all... ¶ 11/23/02 it’s going to be close to that: around 520,000 feet ¶ 01/23/03 It was 4900 records and 597,000 feet!

  Recalculations over the course of production, as indicated by the different dates in this journal entry, show that Murch’s initial estimate for the total number of records is essentially correct. But the amount of footage—nearly 600,000 feet—turns out to be even higher than he projected. “This was the most amount of film for me to handle as an editor on my own, ever.” There was twice as much footage on Apocalypse Now—1,250,000 feet, or just over 230 hours—but on that film Murch was one of three film editors working simultaneously. And post production lasted over two years.

  Technicians at Kodak Bucharest assemble processed negative film before it is transferred to video-tape on a telecine machine.

  The telecine machine at the Kodak lab for transferring film to videotape.

  Why does the quantity matter so much? Is it simply a fact to be marveled at, a statistic to obsess about? Not at all. As Murch later explains, “It’s in my job description. I should be able to tell a director—Anthony in this case—that at this pace of shooting the assembly will be over five hours long.” The length of the assembly matters because it may determine whether the director, who is responsible for delivering a film on time and on budget, fulfills his or her contractual obligations. Other crew members and production executives keep track of production costs and scheduling issues; but only the editor can predict with any certainty if the schedule for editing is accurate, given the amount of work and footage to come. Moreover, the time it takes to edit a motion picture relates, in large part, to the length of the first assembly. The more footage an editor and his or her crew have to begin with, the longer it will take them to assemble it all, and then the longer it will take to pare it all back down to a releasable length. Cold Mountain was originally scheduled for 26 weeks of post production in London. While still in Romania, the producers adjusted this timetable upward; in the end, post production wound up lasting 46 weeks.

  At this early stage, Murch also has his eye on what he calls the “30 percent factor”—a rule of thumb he developed that deals with the relationship between the length of a film and the “core content” of the story. In general, 30 percent of a first assembly can be trimmed away without affecting the essential features of the script: all characters, action, story beats will be preserved and probably, like a good stew, enhanced by the reduction in bulk. But passing beyond the 30 percent barrier can usually be accomplished only by major structural alterations: the reduction or elimination of a character, or whole sequences—removing vital organs rather than trimming fat. “It can be done,” says Murch, “and I have done it on a number of films that turned out well in the end. But it is tricky, and the outcome is not guaranteed—like open-heart surgery. The patient is put at risk, and the further beyond 30 percent you go, the greater the risk.”

  In the case of Cold Mountain, a first assembly of five hours would mean that the 30 percent barrier would be encountered when the running time had been trimmed to three and a half hours: still too long, most likely, for theatrical release. To get below that length, the prognosis is for drastic surgery—unless some solution can be found now, during production, to cut the script or change the pace of shooting—very serious considerations.

  But for now it’s landmark day. The workflows, patches, plug-ins, and workarounds that Sean Cullen and DigitalFilm Tree put together finally pay off: The first Cold Mountain footage pops out at the end of the pipeline, digitized, on Walter’s Final Cut Pro desktop.

  The first shot from Cold Mountain to be digitized—Swimmer (Jay Tavare)—in the battle scene.

  July 24, 2002, Murch’s Journal

  First Digitize!! And we set the rate at two megabytes per second. Looking at a CU of Swimmer. It looks very good. Congratulations, we are beginning to be sort of up and running.

  Once again, there isn’t much time to celebrate. A movie is all about story, and the demands of story soon assume center stage. One of the Miramax executives, Bob Osher, visits the set and sits in on dailies screening. He takes note of the same troubling issue that Murch and the director of photography, John Seale, discussed earlier: will an audience understand how the Union army gets stuck in a pit they themselves created when they set off the massive explosion beneath the Confederates?

  The opening shots of Cold Mountain show Union soldiers laying explosives under the Confederate fortifications. The huge explosion beneath the Confederate army’s entrenchment momentarily buries Inman, the protagonist. The Union army’s next move—a ferocious charge at the disrupted Confederate lines—occurs while a mushroom cloud rains dirt and mud down on all the fighters. It isn’t easy to tell one side from the other, nor is it entirely clear that the Union’s charge into the crater wasn’t deliberate. How will the film be able to show events clearly enough for the audience to understand what is happening, yet not so clearly that the Union soldiers look foolish for trapping themselves? It’s a proper question for Murch to have on his radar, but it won’t truly be addressed until he begins cutting the scene.

  * * *

  Minghella on Filming the Battle Scene

  When we were shooting, there was a moment of it being absolutely overwhelming. We were right down in the crater, hand-holding the cameras with a thousand people milling around us, and for a second I got some sense of what it must be like in a war when you can’t hear, you can’t think, and you can’t see, and you’re just falling. It was 105 degrees, and the actors wore prickling wool uniforms and were just passing out. There’s nowhere to go, nowhere to hide. I thought if I could get it right—obviously I’m not noted for my action sequences—it would be indelible in the film.

  —from an interview in the February/March issue of Written By, published by the Writer’s Guild of America

  * * *

  Meanwhile, Murch’s request to bring on a third film assistant has been approved. Final Cut provides digital film sound that can easily be played back from a portable Akai digital dubber, so location dailies will be screened using a portable Arriflex Loc-Pro projector linked to the Akai—a system that any production assistant can operate. Consequently, the money budgeted to hire a projectionist is freed up to bring Dei Reynolds to Bucharest from London. But Reynolds hasn’t yet arrived and pressure mounts. Being shorthanded and under the gun, Walter works with his son and with Sean Cullen to prep film dailies. The job of ushering film through the lab, rearranging the footage in the order Anthony wants to see it, and getting it synched up with sound on the ProTools system—all in a morning—is usually handled solely by assistants. But as the footage piles up Murch must step in to help. In his journal he simply writes, “Too frantic dailies prep.”

  During the battle s
cenes, the first and second camera units are both shooting at the same time, each with up to four cameras. Plus there are five Eyemo cameras, 60-year-old refurbished hand-held movie cameras originally used during World War II. The amount of film being processed every day averages 5,000 feet, or almost one hour, though on certain days that balloons to two hours or more.

  Inman (Jude Law) buried after the explosion underneath the Confederate’s position at Petersburg.

  Minghella requested that the dailies be broken down into “selects” and “non-selects”—select rolls being made up of just one good take from each camera setup, in shooting order. This gives him the flexibility (after a long day’s shooting) to see a short but essentially complete selection of dailies—20 minutes of film on an average day. But the non-selects (the other forty minutes of dailies) have to be available, just in case. A camera setup is literally that. If the camera is moved—or even a lens changed—for the purpose of getting a closeup angle of what was just shot from a medium angle, for example, that is considered a different setup, and given a new identifying number. The actor’s dialogue and movements may be exactly the same, but the shot is regarded as distinct.