Constrain Mode Limitations and Workflow

< Previous | Contents | Manuals Home | Boris FX | Next >

Constrain Mode Limitations and Workflow

The constrain mode has an important limitation, while initially solving a shot in Automatic solving mode: enough constrained points must be visible on the solving panel’s Begin and End frames to fully constrain the shot in position and orientation. It can not start solving the scene and align it with something it can not see yet, that’s impossible!

SynthEyes tries to pick Begin and End frames where the constrained points are simultaneously visible, but often that’s just not possible when a long shot moves through an environment, such as driving down a road. The error message “Can’t locate satisfactory initial frames” will be produced, and solving will stop.

In such cases, the Constrain mode (checkbox) must be turned off on the solving panel, and a solution will easily be produced, since the alignment will be performed on the completed 3-D tracker positions.

You can now switch to the Refine solving mode, turn on the Constrain checkbox, and have your constraints and pegs enforced rigorously. As long as the constraints aren’t seriously erroneous, this refine stage should be quick and reliable.

Here’s a workflow for complex shots with measured coordinates to be matched:

1. Do the 2-D tracking (supervised or automatic)

2. Set up your constraints (if you have a lot of coordinates, you can read them from a file).

3. Do an initial solve, with Constrain off.

4. Examine the tracker graphs, assess and refine the tracking

5. Examine the constrained points view to look for gross errors between the calculated and measured 3-D locations, which are usually typos, or associating the 3-D data with the wrong 2-D tracker. Correct as necessary.

6. Change the solver to Refine mode

7. Turn on the Constrain checkbox

8. Solve again, verify that it was successful.

9. Turn on the Peg mode for tracker constraints that must be achieved exactly.

10. Solve again

11. Final checks that pegs are pegged, etc.

With this approach, you can use Constrain mode even when constrained trackers are few and far between, and you get a chance to examine the tracking errors (in step 4) before your constraints have had a chance to affect the solution (ie possibly messing it up, making it harder to separate bad tracking from bad constraints.)

Note: if you have survey data that you are matching to a single frame, you must use Seed Points mode, must make each tracker a seed, and you must turn on Constrain.

©2024 Boris FX, Inc. — UNOFFICIAL — Converted from original PDF.