|
||||||||||
|
Anticipate the lighting environment
I'm watching the Shuttle undock from the Mir this morning. At about 7:30 AM CDT, they had a wonderful opportunity to photograph sunrise on the Mir, as seen from the Shuttle. The scene started with running lights outlining the Mir, and then the sun came up.
Sunrise comes fast in orbit, too fast for the Shuttle camera's auto-iris to adjust. So instead of seeing the Mir unveiled inside its running lights, we saw a white blotch leap into view. By the time the camera caught up, the event was over.
We can cope with this by anticipating these events and calculating light levels anticipated.
Masking and merging
In the scenes showing the Mir, the foreground object (Mir) is so bright that the background (stars) is completely washed out. The scene would be much more dramatic if the starfield were visible in the background. This isn't hoke; in views that include only the tail of the Shuttle, the bright stars are clearly visible.
To accomplish this, we need to program the electronics to identify and mask foreground objects in a background camera, while the foreground camera photographs those objects.
We can do this by programming an electronic traveling matt; either by modeling the objects and calculating the matt, or by using the view from the foreground camera to create the matt.
It would be possible to just program the starfield and dub it in to capture what the scene should really look like, but the real picture will carry a lot more meaning. There's an important point in that: we want to capture what space flight is really like. In this case, we're overcoming the limitations of our equipment, not presenting a false impression for the sake of dramatic imagery.
Controllable slew rates
When the cameras tilt, pan, and zoom, they ought to have a programmed acceleration to a slow slew rate. That will prevent the jerky motion we see on the NASA TV images from Atlantis.
We'll need an override so we don't miss the good stuff, even if have to sacrifice the transition shot.
|
|
|