|
Martin Kelly worked for many years as a special effects cameraman at Granada TV in Manchester, England. His work often involved creating "slitscan" effects for title sequences, so he knows more than most about how Douglas Trumbull went about generating the "stargate" scenes in 2001: A Space Odyssey. He has put together this guide for the more technically- challenged among us.
No doubt many people have set their camera shutters to the open position to capture the light trail left by sparklers at bonfire night, or to show the trails left by car headlights on a busy road. The whole path covered by the points of light is captured on a single frame. The camera does not move, but the length of exposure allows an entire cycle of movement to be recorded.
"Slitscan" is a way of creating a similar effect from still images or transparencies. Let's take a simple transparency consisting of bands of colour, which is placed on a light box with a rostrum camera mounted above it (figure 1).
The camera assembly can move vertically up and down, and the platform holding the transparency can move "north", "south", "east" or "west", as well as rotating in either direction. This gives an infinite number of possible positions of camera in relation to transparency (within the limits of the assembly size, and the lens used). To achieve the required degree of precision, the movements are computer controlled.
Now let's take the "slit" - literally, a narrow slit cut out of a mask, wide enough to cover the width of the transparency required for our image - and place it over the transparency (figure 2). The camera, with frame one in the gate, shutter open, tracks in (i.e., down) from the position shown in figure 2 towards the slit, coming to rest in the position shown in figure 3.
At the end of this movement, see how the slit has moved relative to the camera, so it is just out of shot. The shutter is closed; one frame has been exposed.
|
|
|
| Figure 1 | Figure 2 | Figure 3 |
In figure 4, the camera's field of view at the start of the operation is shown by the red outer rim; the inner red rectangle shows the field of view at the end. However, because the movement has been recorded on a single frame, what is actually exposed on film is something like the image in figure 5.
|
|
| Figure 4 | Figure 5 |
A similar process is followed to capture an image on each successive frame. An additional sense of movement towards the horizon is achieved by moving the transparency a little "southward" as the exposure is made - you can see this in the slight difference of position of the transparency in figure 3 compared to figure 2.
Given a motion picture film projection rate of 24 frames per second, filming a complete 10 second slitscan shot would mean repeating the process 240 times. Varying the distance that the transparency moves per frame will increase or decrease the apparent speed.
Of course, the process described above is an extremely simplified form of the highly complex sequences needed for 2001 A Space Odyssey, but the principle is the same. Douglas Trumbull and his colleagues spent many hours creating and selecting the right transparencies to scan to create complicated mixes of colour and shape, and several months putting the whole sequence together. Some of the results in the "Stargate" sequence are still baffling in the extreme, even given a rudimentary understanding of how they were done.
For comparison, here is a photo (figure 6) of the actual Oxberry Animation Stand, or Rostrum Camera, that I used to operate at Granada in the heyday of film. These days they use a Video Rostrum Camera instead, but you can still do some pretty amazing things with the original!
Figure 6
| Back to top | Copyright © Martin Kelly, The Underview, 1999, 2000, 2008 | Back to "How Was It Done?" |