Gordon Willis, ASC Interview at AFI — Part I

A brilliant, often controversial cinematographer shares his considerable expertise with student filmmakers of the American Film Institute.

This interview was originally published in American Cinematographer, September 1978. (Part I of II)

As perhaps the most important aspect of education for the Fellows in training filmmakers, historians and critics at its Center of Advanced Film Studies, located in Beverly Hills, California, the American Film Institute sponsors conferences and seminars with top technicians and talent of the Hollywood film industry. These men and women, outstanding professionals in their respective arts and crafts of the Cinema, donate generously of their time and expertise in order to pass on to the potential cinema professionals of tomorrow the benefits of their vast and valuable experience.

In keeping with this tradition, Cameraman’s Local 659 (IATSE) sponsors a continuing series of seminars with ace cinematographers. These men — both contemporary working directors of photography and some of the now-retired “greats” of the past — meet informally with the Fellows at Greystone, the magnificent estate which is the headquarters of the AFI (West), to present valuable information on cinematographic techniques and answer questions posed to them. Very efficiently introducing and moderating each of the individual seminars is Emmy Award-winning director of photography Howard Schwartz, ASC.

The dialogue which follows has been excerpted from the AFI seminar featuring famed cinematographer Gordon Willis, ASC, whose credits include: Klute, Parallax View, Up the Sandbox, The Paper Chase, The Godfather, The Godfather: Part II, All the President’s Men, Annie Hall and Woody Allen’s latest production, Interiors.

A highly intelligent and articulate artist with impeccable technical credentials, the frequently controversial Willis is considered by some within the motion picture industry to be a kind of maverick, because he does not hesitate to break the “rules” of cinematography in order to gain the effect he considers correct for the project at hand. Some of his refreshingly off-beat views are expressed in the following dialogue, which was preceded by a screening of The Godfather: Part II, on which he functioned as director of photography.

Howard Schwartz, ASC: You all know that Gordon Willis has done a lot to change attitudes toward cinematography — in terms of what is acceptable and what is good. It’s a whole different game that he introduced. I think it started with his work on Klute. He stayed away from lighting from the floor, he lit from up high, he had eye shadows that were natural to the situation – and what was a very daring thing to do at that time. You’ve just seen his work in The Godfather, Part II. For the first Godfather film he didn’t have any light on the walls. This time he didn’t have any light on the walls or the people.

Gordon Willis, ASC: Right. [laughter] You shoot fast working that way.

Can you tell us a little about your professional background?

Well, I was an assistant cameraman, and I was an operator for a while, and then I finally got into features as a cinematographer. But along the way I shot commercials, industrials, documentaries — that kind of thing.

How long did it take you to shoot Godfather II?

Ten months. Actually, there’s quite a difference between the two Godfather movies. If you look at Part I, you’ll see that most of the movie takes place in little rooms. But in Part II we were all over the place. We went from Lake Tahoe down to Los Angeles, to the Dominican Republic, to New York, to Trieste, and then to Sicily. I think I left one place out, but that’s all that I can remember.

What was shot in Trieste?

The opening of the movie — the interiors of Ellis Island. There was a big fish market where we did the immigration scenes and it was very much like Ellis Island in the past. And the people that we used there we could not have gotten in America. Then the shot of the Statue of Liberty I made in New York harbor about a year later. I was the last living person on that picture to make a shot. So the opening is really cut up between New York, Trieste and Rome. There’s a little interior shot of the boy inside a room. We did that in Rome.

The Ellis Island sequence is interesting in that it’s the only interior sequence I’ve seen you do that had the feeling of being over lit which was done on purpose, naturally — in order to get a bleached-out feeling. How did you handle that?

All of the material for the opening Ellis Island sequence was shot at T/2.8, and all of it was one stop overexposed. That’s an oversimplification, but on a mechanical level, that’s what it was. The rest of the material was not shot overexposed. It was what I call “on key.”

“That’s the reason I can’t look at it anymore. There are a lot of things in it that aren’t right, but rather than be reminded of it, I have to close it out of my mind.”

Did you shoot at ASA 200 or 400?

I shot more or less with the normal ASA rating. The contemporary material was a half-stop underexposed — which was normal for the contemporary — but I worked very high on the period material.

And how many nights did you spend at the lab getting that color into it?

Well, unfortunately, The Godfather took eight weeks to print, but we printed Part II in 10 days. Francis Coppola kept making changes. He kept re-cutting. This was the last I-B printing to be done in the United States, with dye-transfer work at Technicolor. But every time you make a cut in an I-B roll you have to redo the matrices. That’s 2,000 feet. So every time he’d make a change they’d have to make new matrices, and he kept cutting right up to the very last minute. I finally got him on the phone and said: “Are you going to make any more cuts? This is impossible. We can’t get it together.” He finally finished, and that gave us a week before the film went into release. So we did the best we could, but I was getting tired of shooting during the day and spending the evenings at Technicolor.

You say that the prints were made by the I-B process. Does that mean that you used separations for all your timing?

Well, to give it to you from the very beginning. I’m a one-light cameraman, I pick a light and a color ratio for the movie, and then ask for one printer at the laboratory. The lab just has to do the same thing every day. They don’t change anything, although I change things back and forth. Now, from those dailies — after everything has been cut — they make separations (which are the matrices) and then finally the print.

Was that how, in the party sequence at Lake Tahoe, you were able to let it go a bit yellowish, while the water stayed blue?

No. All of that was based primarily on the filter pack — the ratio that was used. It was just an overtone of yellow, which I used on Part I, as well — but if you have a second chance to do something, you try to improve on it. I feel that I made improvements in Part II. Also, it was a very sophisticated movie. They jump ahead 15 years in the story, and then jump back to before the first story — so that really, if we’d had the time in the laboratory, there would have been three different tones in the movie. We almost got it, but we ran out of time, as I’ve explained. I make it as simple for the laboratory as possible, so that they can do everything without a lot of changes. The hardest thing to get a lab to do is to leave what you’ve been doing alone. I mean — they feel that they must time the movie, whereas I’ve already timed it from the start. They’re too expensive to use anymore. They may be used in a couple of foreign countries, but no more in the United States.

“Well, I went home and got drunk after I heard that they were ready to punch out a thousand prints of a half-timed movie. That’s when I started screaming and yelling and all that, but it all calmed down.”

Then that’s why Technicolor Plant Four closed?

Yes. They kept it open for us. There was a lot of hysteria, because they’d already had so many advances of money, and where was the movie? They said, “We’ve got to get this out; we’re closing down the plant.” And I won’t tell you what I said, because you can’t print it. But I got Francis on the phone, and we kept the plant open, and got more of the materials needed for printing out of Eastman and finished off the initial run — which in itself was an accomplishment. That’s the reason I can’t look at it anymore. There are a lot of things in it that aren’t right, but rather than be reminded of it, I have to close it out of my mind.

How many prints were made?

Well, I went home and got drunk after I heard that they were ready to punch out a thousand prints of a half-timed movie. That’s when I started screaming and yelling and all that, but it all calmed down. The next crisis came during the exhibition. I’m sure you’ve all heard that when they put movies in the theaters, they’ll just arbitrarily drop out a reel in the middle of the show, so that they can sell more popcorn or get an extra showing in. Evidently they’d been dropping a double reel out of this show here and there. The movie’s complicated enough without dropping 20 minutes out of it.

If you were to reshoot The Godfather: Part II would you go for those low lighting levels again — to the same degree, I mean?

Probably. There’s a philosophy to it — kind of a Greek tragedy thing. I surrounded Pacino with all that because of what the character was and the environment was. I made it even darker than the first movie for that reason. In my opinion, it worked. The only time it doesn’t work is when you run it in a theater with below-standard projection. In order to be at a standard for viewing film, a theater screen and projector must produce 16 foot-lamberts. I mean, that’s it. And if it’s not, you’re not going to see what you’re supposed to see.

You said that you shot most of the footage a half-stop underexposed. Do you get any problems from having such a thin negative — you know, like negative scratching?

Well, I’ve heard all the stories. I get them from the laboratories and producers and all that. But I don’t really have any problem. There’s a translucent quality to exposing at that level — which I like of like. There’s a great deal of latitude in the Eastman color negative. A lot — an incredible amount. But you have to know where it is you’re going to put it.

In regard to your question about scratches on a thin negative, I’d like to say that you underexpose day-for-night two stops and you don’t have a scratch problem, so that’s not a consideration.

Your biggest problem in anything you do — and I don’t mean this to be a negative remark — keeping control at the laboratory level.

Do you underexpose just to keep from getting detail in the shadows and then print up for the release print? Because I would think it would make your flesh tones go too dark.

No. As I said, I pick a printing light for the movie and I work to that light. I don’t print up or down. The dailies looked about like what you saw in this release print. What I mean is that if you A and B’d the dailies with this print, you’d see that they were just about the same.

“I’m very proud of the movie. It may fail on certain levels, but it’s an extraordinary piece of filmmaking, if you look at it carefully.”

This is meant as a compliment, but was there any effects photography in the picture — front projection, rear projection, bluescreen?

There was a quick shot of De Niro on a train in which we used rear projection.

In those flashback sequences to an early period on a New York street, was that a painting in the background?

No. That was the real thing — three blocks of an awful place in New York, the Lower East Side. But it’s the only place left that looks like that. It was pretty dreadful.

About your lighting of certain close-ups — there were several times when the lighting was so low key that the actor was just silhouetted. You couldn’t see his face to see distinctly what he was feeling or anything. I was wondering what you felt about that.

I’ll answer that in a general way, because I’m not a great believer that you have to see an actor all the time on the screen. I believe that the scene has to be played properly, but sometimes it’s better not to see what is going on until a given point in the scene. Then you see something.

Were these things discussed with the director — such as when you wanted to see a face or when you didn’t want to? Did you arrive at an understanding about that somewhere along the line?

Well, I’d worked with Francis on The Godfather: Part I — which was a hair-raising experience — and we had done a lot of lighting on that one. But on Part II we didn’t. We had a good relationship and we had a lot of fun. What happens is that you get to know about each other and you tend not to discuss as much as you would, say with a director who is new to you. Such a director is having a nervous breakdown until the first week of dailies, and you’re worrying about whether he likes this or that. But when you work with a director for the second time, it minimizes the discussion. He knows what he’s buying. In my case, he knows that if I’m going to shoot his movie, he can almost predict how certain things are going to look.

How much did you guys talk about concept on Godfather I?

On this first one we were too busy fighting with Paramount to discuss anything. It was like staying on your feet 24 hours a day. But conceptual discussions were long on Part II and very interesting. I’m very proud of the movie. It may fail on certain levels, but it’s an extraordinary piece of filmmaking, if you look at it carefully.

I’d like to ask about the festa sequence — that parade through the streets of New York. Was that a real festa or was it done for the movie — and did you have to light the whole street?

The festa was stages specifically for the movie. As for “lighting” the street, I’m a bit tongue-tied about that because it was such a nightmare. It’s things like that which sometimes make me want to become a still photographer — because you don’t have to cut anything together. A still photographer doesn’t have to match light from scene to scene. We started the festa sequence in overcast — and I might say here that I think the only thing that Francis and I didn’t really agree on was the use of soft-light. He didn’t like it, but I like it for period work because I think it tends to look more like period lighting. Anyway, trying to shoot that entire festa sequence in overcast turned out to be a horrendous experience. It took weeks, because we kept cutting it up from day to day. The sun would come out and we wouldn’t be able to shoot. I had tarps strung for two blocks (because it was an east-west street, which was swell), but there were windows built all the way along one side, above the period work which had been done. When the sun would hit them, the reflections would come down and hit the street.

“I don’t like blue in movies, especially period movies. I don’t even like it in contemporary movies. I think it’s a vulgar color on the screen.”

The sequence looked very moody. Did you use any smoke or fog filters?

I used a low-contrast filter for all of the period work, which is touchy, because I feel it’s like relying on fog filters. A lot of people’s answer to period work is: “Well, I’ll drop it in a fog filter.” But that’s bull, because it doesn’t do anything. It’s basically the photography that has to carry it. Just dropping something in front of the lens won’t do it. As for smoke, there was one point in the sequence where the fireworks were throwing a lot of smoke around.

Apart from low-contrast and 85 filters, did you use any other diffusion or color filters?

Nothing — except for the required neutral density filters, which meant that the operator couldn’t see half the time. You see, all the period work was shot at T/2.8 and it was all forced-developed, exterior and interior. I don’t have to tell you how horrendous the neutral densities were that I had to drop in, because not only was I shooting at T/2.8, but I was forcing it. That meant that there were a lot of ND filters that the operator had to look through.

Why did you choose to shoot at T/2.8 for daytime exteriors?

Well, lenses change qualities as you stop down, so I just chose that stop for the entire sequence, in order to maintain a consistent feeling.

I’m curious about the yellowish-reddish color cast you had in one scene. It was a period scene in an olive oil office and you were shooting from the back of the office toward the window. Did you intend for it to be that warm?

Yes, I know what you’re getting at, but let me tell you what happened there. I never could straighten it out. As you know, one of the things you have to watch out for when you’re shooting a movie is to maintain lighting continuity, so that the scenes within a sequence will cut together smoothly in the visual sense. There was a lamp in that room over his desk and it was prominent in the master shot, but the master shot was cut out of the movie, so you never saw the lamp. In the closer shots, I had put a color cast on De Niro’s face because of the lamp. So I made a mistake; that’s what it amounted to. I should have either placed in the lamp so that it was always in, or cut it out completely.

In that scene, the exterior showing in the background went a bit blue. Was that because you used no 85 filter?

No. There was an 85 on the camera. How you light an interior-exterior scene depends upon the circumstances. Sometimes you shoot tungsten inside and sometimes you light for a daylight balance, depending upon what is more convenient to work with at that moment. In this case, the lights used inside were of a lower Kelvin. That’s all that made the interior warmer.

I wanted to ask you what type of units you used with blues on them inside that gave you so much control over them. You must have used pretty good-sized units. You didn’t use soft lights, did you?

No. When I did this movie, I’d finally refined a system which I’d been using for several years. What it involves is either daylight photofloods, five to a diffuser — or it might be FAY lights bounced off. I’m using that less and less now because it’s so cumbersome. So I take daylight photofloods and fire them through a diffuser, which is in the ceiling. But then I’ll take quarter booster blues, which are available from Rosco, and I’ll just keep adding them. A daylight photoflood reads 4800 Kelvin anyway, so it’s red to begin with. I either add blue or subtract blue, depending upon what it’s supposed to look like.

The period interior sequences in New York were very, very brown and there was no blue in the scenes at all. Did you influence the art direction in order to get that effect?

Yes, I don’t like blue in movies, especially period movies. I don’t even like it in contemporary movies. I think it’s a vulgar color on the screen. I don’t hate it, I wear it, but I mean that on the screen I don’t like it. I think it overwhelms actors and overwhelms the screen.

“I used the same camera and the same lenses on the second one as I did on the first one — mainly because I’m hopelessly romantic and I thought it would be a nice idea to do that.”

You used a lot of blue in Parallax View, if I’m correct.

There was a lot of blue in the light on that picture. Those were mercury-vapor lamps. But that, again, was a total treatment in a room, as opposed to someone walking around in a blue jacket.

What kind of lighting units did you use outside?

Sometimes I use absolutely nothing, but on Part II I used probably more exterior light than I have in the past, because of the period element. I needed a lot of control in order to make it work. So I used arcs in Sicily. I used arcs at Lake Tahoe — for two reasons. First of all, it’s very expensive to shoot there. Secondly, if you’ve ever been to Tahoe, you’ll understand why the Donner Party got trapped there and all of them ate each other. The weather there is horrendous. In about 20 seconds you’ll have a snowstorm, and in another 30 seconds the grass will be growing. It became a horrendous experience trying to paste exteriors together there. In a case like that, it pays to have lights that can help you out, because you can’t make sunlight. The producer was always coming up to me and asking, “What are we waiting for?” I said, “I’m waiting for you to ask me what we’re waiting for.” He said, “Well, why can’t we shoot?” I said, “Because we have 15 minutes on continuity in the sun, and now it’s raining. We can’t do anything else.” But in general, I used a lot of light primarily for the look.

That snow sequence at Tahoe, where Michael’s inside a room talking to Fredo — was the snow an accident?

The snow was actually an accident. A couple of days before that we were photographing the party sequence outside, with the band and the whole thing, and then it snowed. So we took advantage of that particular snowfall for the scene in the sun parlor. That night the roof collapsed because there was so much snow. It’s wonderful up there. I can hardly wait to go back.

You told me you love snow.

I do, as long as it’s not in the movies I’m working on.

For shooting that period stuff in New York, did you use old lenses — old Cookes, old Baltars?

I used Baltars for both movies. I used the same camera and the same lenses on the second one as I did on the first one — mainly because I’m hopelessly romantic and I thought it would be a nice idea to do that.

I was curious about why you picked the interior-exterior balance that you frequently did. You would let the interior foreground go dark, but what could be seen out a window would be hotter than normal, and bluer, too. Was that concept a matter of your personal taste?

That concept applied to this movie, as far as Lake Tahoe and the boathouse were concerned. As for it being bluer, again that was because the Kelvin of the interior light was lower than the Kelvin of the exterior light.

But I’m sure you had the facility and resources to say, “Put more neutral densities on the windows.” Why did you elect not to?

I always elect not to do something that is going to be more complicated than is necessary. As it was, there were $20,000 of neutral densities that were cut for that boathouse. But whenever anybody elects to look out windows in a movie, the price goes up. I mean on location. I use about a three-stop balance between interior and exterior, because at three stops you can still see people. Now, if you want to see more detail, then you’ve got to start throwing stuff on the windows and bringing up the interior light level and all that kind of thing.

“I’ve never used a color temperature meter. I’m not a Kelvin freak. I know what the standards are and I like to bend them to this or that. So I do it all by eye.”

So your choice was based upon time and money, rather than on what you wanted the result to be?

No. It was a combination of both. Had it been time and money, I wouldn’t have gotten $20,000 worth of neutral densities to put all over the windows. There’s much more of the action that took place in the boathouse that’s been cut out of the movie. So there was a large investment made to shoot scenes in that room you’ll never see. But it was an aesthetic choice to make it look the way it does. My choices are always aesthetic, but after you’ve made that choice, then you have to decide how you’re going to spend money — what’s best, what’s fastest. You know there’s a limit — although there didn’t seem to be a limit on that movie, now that I think of it.

How much footage did you shoot on Part II?

I don’t really know. The assistants were laughing about it in Sicily. I seem to remember a figure of about 900,000 feet, but…

You were Eastman’s biggest customer.

Yes, I mean they loved us at Eastman and Technicolor.

Does anybody know?

Yes, the editors know the total. By the time it reached those proportions, I was very interested in holding my head together to finish the movie at the right level in Sicily — because if you’ve ever traveled in Rome and into Sicily, that alone is enough to put you away.

Whose place was that where you shot in Tahoe?

It was the old Henry J. Kaiser estate. It was wrecked when we found it, you know. It was coming apart. So the art department went in with a quick half-million dollars and fixed it all up. They fixed some of the inside and some of the outside. They put in a lot of grass — you can’t put in a fake lawn, after all — and it was quite beautiful. In fact, Francis was living there. He moved his family in. There were a lot of spaghetti dinners and we had a lot of fun.

Do you use a color temperature meter?

I’ve never used a color temperature meter. I’m not a Kelvin freak. I know what the standards are and I like to bend them to this or that. So I do it all by eye.

But you were talking about photoflood bulbs and throwing in blues and all that. Would you use a color temperature meter then?

I do it by eye — always by eye.

How would you light an actual interior in which you were going to shoot a lot of day action, as well as a lot of night action?

I would rig the room so that I could work both tungsten and daylight-balance lights. And the best way to do that, I’ve found, is to have an interchange system, so that all I have to do is change bulbs. This eliminates constantly having to relight the same situation.

“The first thing I ask myself when I walk into a place of that kind, or of any kind, is: ‘What does it look like with the lights on that are there?’”

I happened to visit with you that day you were shooting in Corcoran and you had a very interesting rig there. The room was about five times as big as this one and you had 10Ks popping down through the diffuser on the top, so that the people in the background were all lit. You just had to light the foreground where the principals were talking.

Yes. There’s a big advantage to using overhead diffusion, once you know how to use it. It’s not indiscriminate. It’s very discriminate. But it takes care of things very nicely.

Is there any problem in fighting mike shadows from overhead lights?

Not if it’s done properly. The sound people probably have less problem with me than they have with anybody. But they have to watch it, because if I’m working on the edge and they get the mike in the wrong way, they won’t see a shadow, but they’ll be chopping a certain percentage of the exposure off. So they have to keep moving the mike around to find their spots.

Do you ever put white paper or card on the floor to reflect up?

Yes, I do it whenever necessary.

I’d like to ask about the film Parallax View. That convention center at the end — the whole place from the catwalk to below had a tremendously even light source, and I was wondering…

Well, I didn’t light that. I used the source lighting that was there. The first thing I ask myself when I walk into a place of that kind, or of any kind, is: “What does it look like with the lights on that are there?” That place happened to be loaded with mercury-vapor lamps. And I thought, “Goody-goody!” Because if I’d actually had to light that place, I’d have gone bananas.

How do you filter for mercury-vapor — and what kind of cast would you get if you shot it unfiltered?

In that case, the light was unfiltered, because I like that kind of blue quality. But I mix light a lot — like fluorescents and tungstens. I just put a 2C filter in for the fluorescents and it shaves just a little bit of the ultraviolet, the awful part of the bluish cast. In other cases, we just change the fluorescents in so that I won’t have to fool with them. Fluorescents are very versatile.

Are there any fluorescents that are about 3200 Kelvin?

Yes, but you can’t get any light out of them. Nothing happens.

“I really don’t believe that you can just photograph a movie; you have to decide what it’s supposed to look like before you can make a decision on how to light it.”

You said before that you had a lot of preliminary discussions with the director about the concepts. What decisions did the two of you have to make?

By the way, that goes on with me in regard to any movie I shoot. Because I really don’t believe that you can just photograph a movie; you have to decide what it’s supposed to look like before you can make a decision on how to light it. In this case we had to decide how to handle the earlier period material. Structure is very important. We had to decide such things as: how we should structure it, how we should frame it, how we should handle it so that there is definition between the various periods of the story. There’s a lot of tableau photography in this movie, especially in the period work; 90 percent of it is tableau.

What do you mean by that?

Well, tableau or proscenium kind of shooting. Tableau is where a scene will play within a frame without making a cut. But it’s also a more formalized way of framing a shot — so that it has, in my opinion, a better period feeling than a three-quarter long shot from the corner, which tends to lack definition.

I believe you said that during your training period you did TV commercials for a while. What, if any, influence did that have on your style?

It had no influence. But it was very good for me as a learning experience, because during the period when I was working in commercials, they were very sophisticated and expensive. Not that that’s the key to anything, but the point is that I had the opportunity to learn a great deal and file it away in my head. I used more technical reference than aesthetic reference; at least, I did in that period.

The reason I asked is that there seems to be a little school of people these days who are coming out of TV commercials and getting into whatever else. Their styles seem to be similar.

I had an advantage over them. Most of the people who learn the business in television commercials and then move on to features are at a disadvantage. Because you can talk to a lot of people who shoot commercials, and they’ll say, “Gee, if I can only get to shoot a feature. I’m going to direct a feature. I’m going to photograph a feature. I’m going to…” Well, that’s swell — only the problem is that when they have their first confrontation with something that runs an hour-and-a-half, they suddenly realize they have a problem. They have to tell a story and tell it well. Fortunately, I grew up in the motion picture business, so I learned about feature movies before I went into commercials. I learned a lot mechanically in commercials, but as far as structure and story were concerned, I learned that before I went to commercials.

This discussion is continued in Part II.