To run a solve, at a minimum, the solver needs to know of: the relationships between the markers and the skeleton; and the root node of the skeleton which will receive the root translation and rotation.
Here are the steps to set up a simple solve:
- Pose the skeleton inside the motion capture data.
- Select a marker, or group of markers, and lastly select the bone they are to drive (the selection process is similar to creating constraints in Maya).
- Click the
shelf icon to indicate that the position of the selected markers are to be used to drive the bone. - Select the root node of the skeleton. This should be located at the hips, and should be the parent of all other joints.
- Click the
shelf icon, then click on “Add” on the window that comes up, to tell the solver that this is the root node of your system.
At this stage, all of the markers that were created and parented to the skeleton (spheres with lines drawn to the parent) should be in exactly the same place as the motion capture data they are being driven by. Now, if you make a small change to the motion capture data, like to move a few frames forward, or select a motion capture marker and move it, we can get the solver to accommodate for that change.
Running the Solver
To run the solver for the current frame, click the 
icon. The solver will search for the pose that brings together the global markers (mocap data) and local markers (parented to the joints). The joints will be posed only, no keys will be saved or modified.
To run the solver for a range of frames, click the 
icon. The solver will solve each frame and set a key for the pose.
These are the basic steps for setting up a solve. From here, the process can be refined to provide a higher fidelity or better looking result by altering the relationship between the skeleton and the motion capture data, and modifying the attributes that affect the way the solve is calculated.