Behind the Seen Read online

  Sound mixing The Godfather: Part II, 1974. Left to right, Mark Berger, Francis Ford Coppola, Walter Murch.

  Murch first worked with Francis Ford Coppola on the motion picture, The Rain People (1969). Both enjoy seeking out the latest technical filmmaking advances, and in some cases, inventing them.

  Murch holds his two Oscar awards for editing and sound mixing The English Patient (1996), directed by Anthony Minghella.

  By 2002, two years later, Final Cut Pro had reached version 3.0. For Murch and Cullen, the prospect of doing an $80 million studio film on software costing $995 is too seductive to disregard, as is the prospect of living in Apple’s friendly interface and object-oriented, cut-and-paste world. With Walter’s encouragement, Sean uses his spare time to do technical research, pore over Apple’s Web site, and play with the software on his laptop. He starts coming across an outfit in West Hollywood called DigitalFilm Tree. “When I did a Google search, DigitalFilm Tree kept coming up—either a reference to them, or to Ramy Katrib, who had spoken at some conference.” Sean talks to Walter again: “There’s these guys in town, DigitalFilm Tree, and I’d like to give them a call. How would you feel about me saying we might do the next show on Final Cut?”

  “Sure,” Walter says. “Let’s see what develops.”

  Office of DigitalFilm Tree, located on the Sunset Strip in Los Angeles. Sean Cullen contacted the company to find out more about using Final Cut Pro.

  The historic Sentinel Building in San Francisco, offices of American Zoetrope, Francis Ford Coppola’s production company. Here, while working on Apocalypse Now Redux (2001) with Murch, Sean Cullen first experimented with Apple’s Final Cut Pro 1.0 to see how it handled film.

  Ramy Katrib had formed DigitalFilm Tree in 1999 expressly to help film editors, as opposed to video editors, learn and properly configure Final Cut Pro. FCP wasn’t originally released with the idea of editing film, so there was a dearth of reliable information and technical support. Katrib quickly found a niche within the post-production community of editors looking for advice and TV and motion picture productions seeking consultation. The information went full circle, since Apple itself came to DFT for FCP development ideas and feedback from users in the field.

  At the beginning of 2002, serious planning for post-production on Cold Mountain hasn’t even started. All Sean and Walter know for sure is that Anthony Minghella is directing; they don’t know where the film is going to be shot or edited. Tom Cruise is actively negotiating to take the role of Inman, which Jude Law will eventually play. Cullen and Murch understand this much: Cold Mountain is going to be a big-budget, high-profile, studio-backed film with a star-driven cast. Characteristic of the motion picture business, there are no deals, contracts, or handshakes. Anthony simply sent Murch an early draft of the script in the summer of 2001, and Walter enthusiastically agreed to edit the film. Still, it was incumbent on Cullen to be careful when contacting people that word about the project didn’t get out.

  When Cullen calls DigitalFilm Tree, Scott Witman picks up the phone. “I’m thinking about doing this film and using Final Cut Pro,” Cullen says. “I want to talk to you guys about what’s possible and what you’ve done before, and just whether or not you think this could even have a chance of working.”

  Witman’s answer might have ended the whole venture right then and there. “Well, we’re pretty busy right now,” he says. “A lot of people are asking for advice. We can’t give out free help any more.”

  Sean assures Witman he has a professional relationship in mind: “No, no, not free. We’d pay you to consult—to find out what’s going on.”

  “In the first conversation,” Sean recalls, “I sort of held back a little bit. I said it’s a large film, that we’d probably shoot half a million feet, and shoot for a number of months—maybe six months. They were starting to get interested. Witman said, ‘We’ll give you a call back.’ I just left my first name and the phone number.”

  With his dark bedroom eyes and customary three-day growth of beard, the founder of DigitalFilm Tree is known among women in the digital film community as “beautiful Ramy.” He first began exploring the unmapped terrain between film and desktop digital video in 1998 while working the night shift at Magic Film & Video Works in Burbank. It was then considered the largest negative-cutting house in the world, where film negative was conformed for shows like E.R., NYPD Blue, and Spin City, along with most of New Line’s feature films. Ramy was the telecine operator, doing film-to-video transfers.

  Ramy Katrib, founder and president of DigitalFilm Tree.

  “When you walked in, the smell of chemicals and cement just overwhelmed you,” says Ramy. “They had ten splicing stations, old German machines with foot pedals, all going click-click, click-click. At the time that I worked there they had 70 employees, all negative cutting.” (As a measure of digital editing’s invasion of the film world, only 15 workers remain.)

  “Negative cutting is the scariest post work I’ve ever seen,” says Ramy. “If you make an error you lose your job. I hated it.” But in retrospect, the experience was fortuitous both for Ramy Katrib and for Walter Murch. “I stopped being scared of film,” says Ramy. “I touched it. I rolled it.”

  Ramy Katrib’s life began far from Hollywood. He was born in Beirut in 1970 of Lebanese-Christian parents. His father was an evangelist, his mother an English teacher. They settled in Loma Linda, east of Los Angeles, in 1975. From there Ramy went to U.C. Riverside and to Columbia College in L.A., where he learned hands-on film crafts. “School was all about learning how to edit, how to shoot; I liked that,” he says. “But everyone in my family was freaking out. They perceived L.A. as Sin City, even though it was only 50 miles away.”

  On off-hours from his job at Magic Film & Video Works, Ramy produced his own documentaries. “Of course I wanted to be a filmmaker, just like every Joe out here,” he says. One project—still unfinished—was about the writer Mardik Martin, who wrote Mean Streets and Raging Bull. The other documentary was about proton treatment—the convergence of nuclear particle physics and medical treatment. For that project, Ramy bought one of the first Canon XL-1 digital video cameras when they came on the market. He was spending his own money, so finding ways to work cheaply and quickly was always a priority. Avid, which sold its low-end machines at $100,000 or more, was out of the question. Having both film and digital video footage to edit, Ramy’s challenge was to find a way to get film footage into a cheap desktop digital video system, such as Adobe’s Premiere, and then get it out accurately. At the time, this was uncharted territory.

  Edvin Mehrabyan, known as “The Finisher” for his top-notch negative cutting.

  Ramy was on the verge of buying a Fast-601 edit system that had Avid-like capabilities but cost $13,000 when he heard about Final Cut Pro. The Mac computer and FCP software together cost less than $4,000.

  Even though Katrib did telecines at Magic Film & Video on a machine using a sophisticated Silicon Graphics computer and operated a DaVinci color-correction workstation costing six figures, he didn’t own a personal computer or even use email at that time—hard to believe, given his desire for taking FCP beyond its domain. But Cullen doesn’t know any of that when he reaches out for help with FCP. All he knows is that Ramy and DigitalFilm Tree are at the nexus where film editing meets Final Cut Pro.

  Ramy calls Sean back and asks what kind of help he needs.

  “Right now, we’re trying to stay low profile,” Sean replies. “We don’t want anyone to say, ‘What are you talking about?’ and short circuit things. We want to keep it quiet. Is that something you can do?”

  “Oh, absolutely,” Ramy replies.

  Sean continues. “It’s Cold Mountain with Anthony Minghella. Walter Murch is the editor.”

  Ramy pauses to catch his breath. “Whoa!”

  By early 2002, Ramy had built DigitalFilm Tree into a considerable force with help from two colleagues at his old telecine job, Henry Santos and Edvin Mehrabyan. During long nights on the grave
yard shift, the three discovered a shared interest in pushing computer-based, non-linear editing software past its original purposes. Mehrabyan, a Russian-Armenian, was supervising all the negative cutting at Magic. He was nicknamed “The Finisher,” says Ramy. “Everyone knew him. People who had botched negative cutting jobs would come to him in a crisis from all around Los Angeles. He could fix negative tears by artificially slicing on the frame line and splicing it himself.”

  Santos started working at Magic a year or so after Ramy. He was 18 then, a kid out of high school, but he already had excellent skills as a colorist—the person who corrects hues and tones in film and video transfers. “He was a sharp kid,” says Ramy, “so I brought him in as an apprentice to work in telecine.” Ramy already was demonstrating the kind of entrepreneurial, team-building instincts that would later serve him well in making DigitalFilm Tree a reality.

  The next principal to join DFT was Tim Serda, a Macromedia certification engineer who worked on Final Cut Pro, which then had the code name Key Grip. Serda had moved to Apple when it acquired the application from Macromedia.

  Once he had his own Final Cut Pro software, Ramy brought his burgeoning brain trust together to see if they could force FCP to accurately cut film. An essential element—the missing link—was FilmLogic, an application for transplanting video’s 30 frame-per-second databases and edit lists into the realm of 24 frame film. Ramy located Loran Kary, the father of FilmLogic, and invited him to Los Angeles to join in the experiment. Without telling anyone, they all gathered at Magic Film & Video Works one weekend.

  Loran Kary developed FilmLogic, a database for taking 30 frame-per-second video information and converting it into 24 frame film. Apple later acquired the program and renamed it CinemaTools.

  Ramy had finagled Sony into lending him a $15,000 DVCAM DSR 2000 digital video deck for a week. “We did something really simple—we transferred film dailies to a DV format, which was unprecedented,” he said. “For us, that was like—whoa! No one was doing this. I took that tape home and put it my Final Cut system. I put it in my little DVCAM deck, lined it up, captured it. And it behaved the same way that video behaved on an Avid. I knew we were onto something. It was too dramatic. And the fact that I was even approaching a place where I’m doing something with film that wasn’t on an Avid Film Composer—I was just smart enough to know that this was big.”

  The interface for FilmLogic, the application used in conjunction with Final Cut Pro to convert information for 30 frame-per-second digital video to 24 frame-per-second film.

  The group’s accomplishment was considerable for two reasons: first, Final Cut Pro software had been tricked into doing something Apple never had in mind when it designed and released the product solely for video; and second, an affordable, open-format, user-friendly competitor to Avid was being hatched in front of their eyes. Within a year Apple caught on to the implications, bought FilmLogic, renamed it Cinema Tools, and hired its creator, Loran Kary.

  “The day we documented that test was when we launched DigitalFilm Tree,” says Ramy. “I don’t know why I decided to document it, because really, if you go back to that day, no one thought anything about anything; it was just kind of hanging out. I even remember a little grumbling from some of the people because it was a 16-hour day. It was no money, no nothing. But we cut a sequence in my apartment, where my system was. We generated a cut list out of FilmLogic. We went back to Magic, and Edvin cut it. And it lined up. It was just like an Avid cut. It was perfect; and everyone was there to verify it.”

  A test organized by Ramy Katrib in 2000 that proved Final Cut Pro, with FilmLogic, could accurately cut film.

  Katrib, Edvin Mehrabyan, and Loran Kary prepare to begin the test in Katrib’s apartment.

  “Once we did that test, I became an evangelist, like my dad,” says Ramy. “I was preaching it to everybody. And that’s when I started talking to Edvin and Henry—telling them this was going to change the whole landscape for film. We went through the process of hiring a lawyer, formed an LLC, and started DigitalFilm Tree.”

  It’s nearly two years later when Sean Cullen drives out onto Santa Monica Boulevard from The Lot. He heads west toward Sunset Boulevard and finds DigitalFilm Tree’s home: a two-story English Tudor-style house, cater-corner from Larry Flynt’s Hustler Store. From the outside, the place looks like the location for Steve Martin’s failing production company in the film Bowfinger—a long-neglected, stereotypically noir Hollywood building. But once he goes inside, Sean sees the future of feature film editing.

  Video monitors show a sample scene being cut in Final Cut Pro.

  Sean recalls that first encounter: “I outlined not only what I thought we might be doing, but some of the politics of Cold Mountain and a description of Walter’s personality. I told them he will say, ‘Oh, this is wrong. It’s off a few frames.’ I said, ‘You really want to be careful, because he’s always right. And you might say, ‘No, that’s in sync.’ And he’ll go, ‘No, it’s out of sync one frame early. Fix it, and then give me a call.’ And sure enough, it’s out of sync one frame early. I said, ‘Watch out, because Walter has a really sharp eye.’”

  “The more we talked,” Sean says, “the more we realized we were perfect for each other. They were doing what nobody else was doing. They were groundbreaking because they were doing the first features. It was clear this was where the smart people were.” What had begun with DigitalFilm Tree saying, “We can’t give out free help,” quickly becomes “We’ll figure out the money later.”

  Loran Kary uses a loupe to inspect cut camera negative—the end product of the successful Final Cut Pro test. Mehrabyan is on the left.

  Cullen describes the sort of relationship he and Murch want to have with DFT. Changing editing platforms means they would need all kinds of troubleshooting and problem solving, especially since Final Cut Pro has never been used to edit a major feature film with so much footage. They will need to be in constant communication, night and day; the inevitable fire alarms will demand an instantaneous response. Sean needs to know if Ramy and DFT can provide that kind of backup and go the distance for well over a year. Barely containing his enthusiasm, Ramy says they can and will.

  Cullen then talks about possible deal breakers. “There are a number of things that, unless they are provided to Walter, he isn’t going to do the film on Final Cut.” Sean tells DFT, “If we can’t make these things work—either because I don’t understand how they work, or they can’t work—then the deal is off, and we’ll do the show on Avid.’”

  “Guys, the stuff on the right-hand monitor looks like crap.”

  One requirement is being able to watch the same material on the computer screen and on the TV screen. By this, Cullen means that in addition to the normal computer monitor, Murch has to be able to use a large-screen TV in his edit room for viewing completed edits. Cullen also insists that both the computer and TV monitors be capable of displaying the same images, simultaneously, cleanly, and crisply by running digitized material at true 30 frames per second. Being a surrogate for the audience, the film editor needs to be immersed in the movie, free from technical distractions, even if he watches a single scene less than a minute long.

  “I knew that Final Cut, when you took it out of the box, couldn’t support a reverse telecine process,” Cullen says later. “You couldn’t watch 30 frames per second on the TV. Avid does this, so I knew that we were going to have to do it. Walter needed to be able to have the record monitor and the TV both play back in real time, simultaneously, not switching one from the other.”

  Ramy takes all these requirements in stride, telling Cullen they can be realized. To prove it, Ramy turns to their Final Cut Pro setup and runs some sample footage. Cullen isn’t impressed.

  “Guys, the stuff on the right-hand monitor looks like crap,” he says.

  “What do you mean?”

  “Walter won’t go for that. That looks bad. It’s jumpy and tearing from the interlace.” For Cullen, the image is compromised
and contains artifacts—lines that are jagged instead of straight—due to the low data rate by which media is being sent to the computer.

  But Ramy sees the glass as half full. “Well, it’s not that bad,” he says.

  “Okay, wait, stop,” says Sean. “We need to understand each other. If I say Walter isn’t going to go for it, he’s not going to go for it. And that isn’t good enough for us.”

  “Well, we know we can get it better,” says Ramy, “just not on this machine.” Like many functions in the Final Cut Pro system, the data rate can be customized—in this case upward.

  Sean is still dubious. “Okay, but I guarantee when Walter comes in, he’s going to need to see it looking really good on both the computer and the TV.”

  Ramy later says, “It didn’t take very long to realize that Sean was a force of nature, a heavyweight. Sometimes when you talk to people, you can feel that some of it is registering and some of it isn’t. With Sean, everything was registering and he was coming back for more.”

  One hugely threatening issue remains, and Sean, Ramy, and the others at DFT talk it through: the fact that Final Cut can create an Edit Decision List (EDL) for conforming the film workprint but cannot track subsequent changes to that list—an essential procedure in any big-budget feature.

  A cut list, or EDL, is the blueprint for how scenes of a film finally get put together. It lists the shots to be used, in order, with each beginning and end point, measured in feet and frames. All effects such as dissolves or fades are also noted. At the end of the editing, when the picture is “locked,” or declared finished, this cut list goes to the negative cutter. Up to that point the film that was originally exposed on the set has remained untouched, except to make the workprint. It’s been sealed in cans, safely locked away in the film lab vault. There is only one original camera negative, and it can never be replaced. Only when all the decision making on the picture editing is done, does the final negative cutting process begin.