< Previous | Contents | Manuals Home | Boris FX | Next >
Though our story is far-fetched, it is quite a bit more accurate than you might think. Though we’ll skip the hundred marching feet, you will be telling SynthEyes exactly how to position the model within the coordinate system.
And importantly, if you don’t give SynthEyes enough information about how to position the model, SynthEyes will take advantage of the lack of information: it will do whatever it finds convenient for it, which rarely will be convenient for you. If you give SynthEyes conflicting information, you will get an averaged answer—but if the information is sufficiently conflicting, it might take a long time to provide a result, or even throw up its hands and generate a result that does not satisfy any of the constraints very well.
There are a variety of methods for setting up the coordinates, which we will discuss in following sections:
Using the automatic Place tool
Configuring trackers individually
Aligning to an existing mesh
Using Phases (not in Intro version)
The three-point method is the recommended approach, as it quickly produces the most controlled and accurate results. The Place tool quickly guesses at something usable automatically, but the setup it produces has no specific relationship to what you want, nor is it accurate: without care, anything you add will slide, and it is your fault, not ours! (That's true in general, actually.) Manual alignment is slow and generally not very accurate, though it allows you to get what you ask for, for sure.
The alignment line approach is used for tripod-mode and even single-frame lock- off shots. The constrained camera path methods ( for experts!) are used when you have prior knowledge of how the shot was obtained from on-set measurements.
You must decide what you want! If the shot has a floor and you have trackers on the floor, you probably want those trackers to be on the floor in your chosen coordinate system. Your choice will depend on what you are planning to do later in your animation or compositing package. It is very important to realize: the coordinate system is what YOU want to make your job easier. There is no correct answer, there is no coordinate system that SynthEyes should be picking if only it was somehow smarter…They are all the same. The coordinate measuring machine is happy to measure your scene for you, no matter where you put it! You don’t need to set a coordinate system up, if you don’t want to, and SynthEyes will plough ahead happily.
But picking one will usually make inserting effects later on easier. You can do it either after tracking and before solving, or after solving.
Hint : if you will be exporting to a compositing package, they often measure everything, including 3-D coordinates, in terms of pixels, not inches, meters, etc. Be sure to pick sizes for the scene that will work well in pixels. While you might scale a scene for an actor 2m tall, if you export to a compositor and the actor is two pixels tall that will rarely make sense.
After you set up a coordinate system and re-solve the scene, it is a good idea to check that everything went OK using the Constrained Points View. Each constraint you added will be listed, along with the error: the difference between what you are asking for, and SynthEyes was able to give you. Normally the error values in the right-hand column should be zero, or very small compared to the size of the scene. If there are large errors, it indicates that the constraints are self-conflicting, ie that you are (often indirectly) telling something to be in two different locations simultaneously.
©2024 Boris FX, Inc. — UNOFFICIAL — Converted from original PDF.