M: Yes, it was a real detective story, involving a forgotten, broken sound cylinder found at Edison's lab in Menlo Park. It started when Patrick Loughney, the head of film and television at the Library of Congress, developed an intuition that this cylinder, which first appeared mislabelled in a 1960s inventory, might actually be the soundtrack for a seventeen-second Kinetoscope that Edison made in 1894. The film is of Dickson playing a violin into a huge recording horn, and it's clear from looking at the image that they must have been recording the sound as they were filming. But the accompanying soundtrack had been lost—some wondered if it had ever even existed. That is, until a few years ago, when Patrick located this particular broken cylinder and had it repaired.
The first known recording of film with sound: frames from a seventeen-second Kinetoscope made by Edison in 1894 showing William Dickson playing a violin into a large recording horn. Murch was able to put the film and sound track in sync.
And it turned out in fact to be a recording of someone playing the violin, with many stops and starts, and fragments of muffled conversation. But the Library of Congress had no means to put the image and the sound in sync: the film was shot at forty frames a second (rather than our standard today of twenty-four) and lasted seventeen seconds, whereas the sound on the cylinder was two and a half minutes long. So the question was: Which seventeen seconds of sound went with the film? And then, once you decide that, how do you put it in sync with the film, which is playing at a nonstandard frame rate?
O: How did you get involved in all this?
M: I was put in touch with Patrick through Rick Schmidlin, who had produced the restoration of Orson Welles's Touch of Evil, and Patrick asked if I could help. I had my assistant, Sean Cullen, digitize both the sound and the picture, and consequently was able to render the film at normal speed and then find various sync points with the music. I tried dozens and dozens over a period of a couple of hours, until I found the one that worked. The soundtrack and the picture were finally in sync with each other for the first time in 106 years!
O: So this is the first known recording of film with sound?
M: Yes. It pushes back the threshold of film sound by a couple of decades. There's anecdotal evidence of something done a few years earlier, in 1891, but neither the film nor the image for that has turned up yet.
As for this fragment from 1894, I'm anxious to use some new technology that's just been developed to see if we can eliminate the surface noise and reveal what might be an off-the-cuff conversation between Edison and Dickson. This would be doubly fascinating! Not only would it be Edison and Dickson, pioneers of cinema, but it would be the first recording ever of a natural conversation between people unaware that they were being recorded.
There's a formality to all recordings of the human voice we have from that period, very much like the photographs of people sitting in their Sunday best, looking right at the camera with blank expressions on their faces. Photography for most people was a once-in-a-lifetime experience: when they looked into the lens of the camera, they were looking at eternity—this is how they would be seen by future generations. It was the same with sound: the experience of having your voice recorded was a magical thing. So you spoke in your Sunday-best voice, very clearly, into the recording horn.
But on parts of this particular cylinder—the violin test from 1894—people were simply talking, naturally. I hope we can decipher it: not only for what they say, but how they are saying it.
O: I hope we don't hear them saying, “He'd kill us if he got the chance.”
M: They probably were….
O: I remember seeing the first filmed interview, done in a mining town in England. One of the first documentaries made, produced, I think, by John Grierson, of a couple whose house is full of rats. The wife is speaking to the camera and the husband is also—like this—looking at the camera, kind of caught in the glare. For the first minute he's just staring at the camera, and then slowly we begin to witness him listening to what his wife is saying, and his head turns and he watches and listens to her in amazement as she tells a story of how she caught three rats. He's completely forgotten the camera … for the first time.
Opposite: From Housing Problems, 1935, a documentary produced by John Grierson about working-class housing problems, in this case, one large rat. Small and big films: Coppola, above left, rehearsing a scene from his second film, You're a Big Boy Now, with Elizabeth Hartman, Peter Kastner, and other actors; and, in striped sweater, above right, under the dolly, checking a shot in Godfather II.
MURDER MUSIC
O: You and Francis Coppola and George Lucas were originally only involved with low-budget/independent films. Did you feel you'd stepped into another kind of realm when you worked on the sound for the first Godfather film for a large studio?
M: Absolutely. It was terrifying! We were in our twenties, remember. It was still us ex–film students, but suddenly there was now the added element of the studio. For a while we were allowed to continue to work in San Francisco. But in October of 1971, Bob Evans—who was head of Paramount—ordered that the film come down to Los Angeles. For the first time I was going to be working in Los Angeles, on a feature film, at a studio, with access to their sound library, but also with all the other things that go along with studio productions, both good and bad.
I learned a lot, not least how much I, even with my lack of experience, could push the studio system to produce unusual results. So rather than simply saying, This is a studio film, this is the way they do it, I kept trying to push the soundtrack in the direction of the films we had already made.
In order to get the right sound atmosphere for the wedding scene in The Godfather,Murch used the “music guide track,” usually thrown away, and mixed it with the original music. He also did this the next year with American Graffiti.
For instance, I was already fascinated with the ability to shift perspective in sound, particularly in something like the wedding scene in The Godfather—the shift from the noise of the wedding outside to the noise of it inside the room, and even when you're outside, to hear different perspectives. I wound up using the music guide track—something that's shot at the time of filming, and usually thrown away—that Francis was playing during the shoot to get people to be in the mood of a happy Italian wedding.
I thought, This sounds great, why throw it away? It sounds like a real wedding. In fact, you could see the loudspeakers in the film. So I painfully reconstructed, out of bits and pieces, a master track of this atmosphere—which had voices in it but mainly that loudspeakery music—then found the original recording of the music and put those two things in sync with each other, so that in the mix we could blend from one to the other and have lots of reverberation and rackety voices, or have it be more up front, with a fuller sound. This was something I also did the next year, on an even more developed level, with American Graffiti. But there are actually little fragments of this technique, in embryonic form, in THX 1138 and The Rain People.
O: This was not the usual way of doing this on a studio film at the time?
M: Don't make me laugh! However, many years later, I learned this was exactly what Orson Welles had done on Touch of Evil, in 1958. He had the idea to use the “bad” sound of recorded music coming over loudspeakers to give a sense of place and as a way of scoring the film using source music.
My twist on the idea was having a track of the good sound too and then being able to fade between both, to have different proportions of one and the other, depending on where you wanted to be. If you were close to the band, you featured the live sound.
Explaining what I wanted to do, and getting that through the machine of the studio, was a challenge. But the mixers actually got excited by it. They thought it was a good idea.
O: Did you have control during the mix of The Godfather?
M: To a certain extent, I did. Francis was directing Private Lives for American Conservatory Theater, up here in San Francisco, so I was his “man in Havana”—I was the
person representing the intentions of the director, which, because Francis trusted me, were frequently my own intentions.
O: Were there any scenes that created problems for the studio?
M: There was an intense crisis with the music. When Bob Evans heard Nino Rota's music, he felt it would sink the film, that it was too lugubrious and didn't have enough energy. He wanted Henry Mancini to rescue the film and to make it more hard-boiled. He didn't like these rather soft-edged ideas that Francis and Nino had come up with—he wanted it to be more American and punchy. So there was a big struggle between Francis and Evans, during which Francis at one point said he would quit the film and take his name off it if that happened.
O: You mean the main theme music?
M: Yes … well, all the music.
O: My God, it's a trademark!
M: Well, nobody knew that at the time. Remember, someone at MGM wanted to cut “Over the Rainbow” from The Wizard of Oz. Frequently what happens in film is that people, especially distracted executives, will say, I hate—pick one— the music, camerawork, art direction, acting in your film. But if you actually get under the skin of that prejudice, you can discover the particular thing they really hate—the pea under the mattress. It often comes down to one or two small things that spoil everything else. When I talked to Bob Evans, it turned out he hated the music for the horse's-head scene, where Woltz pulls the sheet back and the severed head of his half-million-dollar horse is revealed in the bed. Maybe because Woltz is the head of a studio and Evans was the head of a studio and it's a particularly striking, grisly scene—the first violence in the film—he felt the music should be appropriate to that.
I tried to listen to what Nino had written with Bob Evans's ears, and I thought he had a point. The music, as it was originally written, was a waltz and it played against the horror of the event. It was sweet carousel music. You were seeing those horrible images, but the music was counterpointing the horror of the visuals. Perhaps it needed to be crazier a little earlier.
So I tried something I had done on THX 1138—layering the music, playing records backwards, turning them upside down, slowing them down—a version of what I'd done when I was eleven years old.
Nino's music for the horse's-head scene had an A, B, A musical structure. That is to say, it had an opening statement, then a variation, and then a return to the opening statement. This structure allowed me to make a duplicate of the music, slip the sync of the second copy one whole musical statement, and then superimpose them together. The music started off A, as it was written, but then became A + B, simultaneously, and then B + A. You now heard, superimposed on each other, things that were supposed to be separate in time. So it starts off as the same piece of music, but then begins—just as Woltz realizes that something is wrong—to grate against itself. There is now a disorienting madness to the music that builds and builds to the moment when Woltz finally pulls the sheet back.
Murch mixing The Godfather II, 1974.
O: This happened in the shot where we're coming into the bedroom at night, or first thing at dawn?
M: Early morning. All is normal until he starts to stir and realizes that something is in bed with him—that's when this madness, this second element, comes in. Really a replay of the opening statement but harmonically interweaving with the second … You know the way you feel when you're woken by something, and something is wrong, and you wonder, Is something wrong? What is wrong? Oh my God! No, it can't be! It's even worse than I thought! Aaarghhh!
We played this version for Evans, and he thought it was fantastic. He asked us to rewind it as he phoned up Charlie Bluhdorn, who was head of Gulf + Western, in New York. He took the telephone all the way up to the screen and said, “Listen to this, Charlie! Roll it!” holding the phone up to the screen as the music played. I can't imagine what it must have sounded like at the other end of the line, or what Bluhdorn thought, but Evans was very happy: he felt some corner had been turned.
I was sitting at the mixing desk, with Dick Portman, who was the lead mixer, and the whole thing had made Dick very nervous. To do something like this with the music was … well, people didn't do this. It was certainly very risky if it hadn't worked out. So we were both sitting there, looking at this wonderful scene—a big projection in black-and-white of the head of this fictional studio boss discovering a horse's head in his bed as the real head of Paramount stood holding the telephone receiver up against the screen, his shadow cast across it as the scene unwinds. It was one of those iconic moments that you can't believe as it's happening.
The result was that some of the heat was taken off the music. There were still struggles, but they were within the context of Nino's music, not about hiring somebody else to come in and rewrite everything.
O: Was that kind of dissonance used elsewhere in the film?
M: No, that was the only example. It's a very particular scene. The general tendency in The Godfather is to play big scenes in silence and then to bring the music in afterwards. For instance, the killing of Carlo, Michael's brother-inlaw, at the end of the film has no accompanying music. In a so-called normal film you would have dramatic murder music, but we had only that sound of Carlo's feet squeaking on the windshield as he's being choked to death. Then his foot smashes the glass and you're left with the image of his foot sticking through the windshield and the sound of the gravel crunching as Michael walks back to the house. Then the music comes in.
Also the killing of Luca Brasi and even the killing of Michael's first wife in Italy—all these things happen with just sound effects. Music only comes in after the fact.
O: The murder in the car, the shoes on the windshield … I remember talking to someone after I'd seen it about how unpornographic it was, there was no suggestive manipulation, no thrill to it….
M: A fixed camera. The camera is on the hood of the car and it doesn't cut away. You see everything, all through the whole process. Not having music also gives the emotional effect of not cutting away.
When music makes an entrance in a film there's the emotional equivalent of a cutaway. Music functions as an emulsifier that allows you to dissolve a certain emotion and take it in a certain direction. When there's no music, the film-makers are standing back saying, simply, Look at this. Without appearing to comment.
O: Do you know of examples in, say, early gangster films, where the same kind of thing happens—a restraint in the music—or was there always a traditional musical stamp?
M: I don't know enough about film history….
FIVE TYPES OF AMBIGUITY
O: It seems the sound mix is the crucial stage in a film, where everything jells and in some way the film almost doubles in power? Has the process of doing a sound mix changed much since the early days when you worked on The Godfather?
M: The mix is still really the final stage at which any last opportunity can be seized or any last insoluble problem solved. If you're lucky, and if you have the right approach, a certain blend of music and sound can sometimes solve problems that could not be solved in any other way. That's part of the filmmaking process. Every stage leaves a residue of unsolved problems for the next stage— partly because the particular dilemma you're facing cannot be solved in terms of the medium that you're working in right then. For instance, at the script stage there may be issues that have to be left undecided, so the actors can have a fruitful ambiguity to work with. It would be deadly if you did solve all the problems in the script—you do not want to be asking for the gods' help at every stage—because then everything subsequent would be a mechanical working out of an already established form.
The acting, the shooting, the editing, and the sound may all blend into one another, but in fact there are five stages in a film's life: the script stage; the preproduction stage, where you cast and choose locations; the shooting; the editing; and then the sound and music stage. Each is fateful in its own way.
But because the sound mix is the very final stage—and because it's very flexible—there's a tremendous am
ount of variety you can call upon during the mix, by both eliminating things you thought were absolutely essential or, at the last minute, bringing some new element in.
To answer your original question: I guess the thing that's changed since the days of The Godfather is that we now do much more preliminary mixing, what's called temp mixing. This is so we can preview the film earlier than we ever would have thought of doing before. Something Fred Zinnemann, for example, would have abhorred.
The result is that there are generally fewer surprises in the final mix. This can be both good and bad.
O: It's an odd thing: I've heard you talk before about the importance of ambiguity in film, and the need to save that ambiguous quality which exists in a book or painting, and which you think a film does not often have. And at the same time in a mix you are trying to “perfect” that ambiguity.
M: I know. It's a paradox. And one of the most fruitful paradoxes, I think, is that even when the film is finished, there should be unsolved problems. Because there's another stage, beyond the finished film: when the audience views it. You want the audience to be co-conspirators in the creation of this work, just as much as the editor or the mixers or the cameraman or the actors are. If by some chemistry you actually did remove all ambiguity in the final mix—even though it had been ambiguous up to that point—I think you would do the film a disservice. But the paradox is that you have to approach every problem as if it's desperately important to solve it. You can't say, I don't want to solve this because it's got to be ambiguous. If you do that, then there's a sort of haemorrhaging of the organism.
O: And more of a confusion.
M: Yes. I keep thinking about it, and it's a wonderful dilemma: you have to acknowledge that there must be unsolved problems at each stage. As hard as you work, you must have this secret, unspoken hope that one very significant problem will remain unsolved. But you never know what that is until the film is done. You can almost define a film by the problem it poses, that it can't answer itself, that it then asks the audience to solve.