< Previous | Contents | Manuals Home | Boris FX | Next >
24/7 Perspective
With stereo, there are two different views all the time, and even a single frame from each camera is enough to produce a 3-D solve. At least in theory, you never have to worry about “tripod shots” that do not produce 3-D. Every shot can produce 3-D. Every stereo shot can also be used in a motion-capture setup to produce a separate path for even a single moving feature. That’s clearly good news.
But before you get too excited about that, recall that in a stereo camera rig, the cameras are usually under 10 cm apart. Compare that to a dolly shot or crane shot with several meters of motion to produce perspective. And each of the hundreds of frames in a typical moving-camera shot produces additional data to help produce a more accurate solution.
So, even though you can produce 3-D from a very short stereo shot, the information will not be very accurate (that’s the math, not a software issue), and longer shots with a moving camera will always help produce better-quality 3-D data.
On a short shot with no camera rig translation (with the rig on a tripod), you can get 3-D solves for features near to the camera(s). Features that are far from the cameras must still be configured as “Far” to SynthEyes, meaning that no 3-D depth can be determined. Similarly, for motion-capture points, accuracy in depth will degrade as the points move away from the camera. The exact definition of “far” depends on the resolution and field of view of the cameras, you might consider something far if it is several hundred times the inter-ocular distance from the camera.
©2024 Boris FX, Inc. — UNOFFICIAL — Converted from original PDF.