One of the key periods of James Eads' life was the time he spent as a salvager on the Mississippi River. It was during those years that he came to know the river better than perhaps anyone else during the entire 19th century. Recreating his diving experiences on film, however, was challenging.
Director Carl Charlson wanted three shots: A view from the river bed of the diving bell descending from above; a head-on view of the diving bell as it is lowered through the water, and a view of James Eads' hands reaching through the murky water.
Art Director Katha Seidman created a six-inch high model of Eads' diving bell using basswood, lead fishing weights, copper wire and a fish tank hose. The shot of the diving bell from below was achieved by shooting the reflection of the model in a mirror as the model was lowered into a water tank. The camera was on the outside of the tank. The mirror was positioned on the bottom of the tank at a 25-degree angle to the floor.
Cinematographer Boyd Estus explained how those scenes were shot.
What look were you trying to achieve?
James Eads was diving at great depths, in murky water that verged on almost total darkness. What we wanted to do was suggest the murkiness without making an image so dark that the viewer could not see anything. We also needed to make the water flow rapidly to create the sense of the current in the river.
How did you go about shooting these scenes?
Ideally we would have shot a full-scale replica underwater. But we didn't have the budget to do that. A miniature is a fairly economical way of getting the same result. I started out by reading everything I could about shooting miniatures. I have several different editions of the "American Cinematographers Manual" and I read the chapters about history and technique of shooting miniatures in "Special Effects," which date back to 1931. I also spoke to David Quaid, ASC, a director of photography who has worked on many features and is always a tremendous source of information.
The interesting thing is that miniatures were very big in the '30s and '40s (movies like "King Kong") and then people stopped shooting them until "Star Wars." Now we have a whole new generation of people shooting models at a much more sophisticated level. If you look at these early efforts, the miniatures are often pretty obvious due to the technical limitations of the time.
Anyway, what I learned is there are certain rules about shooting models that have been figured out by trial and error since the first films were made 100 years ago.
The first problem is depth of field. When you see a real life-sized diving bell the whole thing is in focus. When I was filming the miniature I was very close, shooting on a long lens, which diminished my depth of field, so I had to compensate by using a lot of light. We were using 2000-watt lights and a lot of them. That's a lot of light in a small space. At the same time we had to achieve the effect that it was dark and with very little visibility. It was difficult to achieve that balance.
Shooting small means you also have to alter the frame rate [the speed the film moves through the camera.] Overcranking creates slow motion and tricks the viewer into believing the object is bigger than it is, that the water has the weight and movement of a large body of water and not a small tank of water. There are formulas that tell you what speed you should be shooting at. The smaller the scale of the model the higher the frame rate. The usual frame rate for film is 24 frames per second -- we shot at 64 fps and 98 fps, which required up to 4 times as much light. Another factor is the movement of the miniature must be appropriate for its apparent size. The formula linking these factors is:
You also need to use filtration to change the image quality for a number of reasons. Underwater your vision isn't that good -- it's not sharp, so you want to create a softer image. Also, when you're shooting a miniature you want to soften the focus to hide the imperfections of the model. And if you are a shooting an object that is supposed to be 30 feet away, in reality there's a lot of stuff in the air or water between you and it that adds diffusion. When shooting a miniature you create this sense of diffusion using special filters.
Finally, we created some water movement, which, with the light reflections, gave the sense of water to the scene.
Were there things you tried that didn't work?
Well, my initial thought was to use a big circular tank because we could put a mixing blade in it that would get the water moving around. Lots of circular tanks are used to create tornado effects. It turned out that circular tanks are hard to come by and very expensive to make. We were fortunate to find a professor at the Massachusetts Institute of Technology who had two tanks that she could let us use. One was about 30 foot long and had a pump that circulated the water. It was perfect for creating rushing water. Exactly what we needed.
What were some of the hiccups in production?
To create the murky water we added cement dye to the tank. As it turns out a little cement dye goes a long way. We had to drain the tank several times because we put too much dye in the water initially. The water became so dark that we could not see anything in it. It took half an hour to drain the tank and another half an hour to fill it up again.
We also had problems with the air coming out of the diving bell. We felt we needed the bubbles to give the image life, to make it seem like there was a human in the diving bell. We were pumping air into the bell through a fish tank hose. But it was hard to get the speed and size of the bubbles right. We needed small, even bubbles. Big bubbles would have given away the size of the bell.
We also had to support the bell in a way that we didn't see the supporting apparatus. That made me think of using a silhouette cutout of the bottom of the salvage boat to hide some of our rigging. We suspended the cutout on the water-- it ended up adding to the reality of the scene.
At one point Katha [Seidman, the Art Director] and I were we thinking of seeing a person in the diving bell. She felt that wasn't practical. She had wanted to make the diving bell bigger -- it would have been easier and more accurate for her to make a bigger model. I insisted it had to be smaller because we didn't have very much tank space. Initially I wanted a tank that was six or eight feet deep, but the one we had was actually only 30 inches deep.
As usual, we sorted it out. Much of the shooting time was taken up in making the countless adjustments and alterations needed to finesse the shots and make the viewer think that he/she was seeing the real thing.