Our workflow for the motion capture sessions was to:

(before the session)

(1)
Select the sequences from the storyboard that needed to be captured.

(2)
Break these down into sections identifying shots that needed to be captured continuously and those that could be captured individually.

(4)
Confirm that the character rigs to be used for the session are the FINAL versions and that their forward kinematics rigs have been completed, tested, imported into MotionBuilder and characterized. Although it is possible to copy animation between different characters in MotionBuilder, it is an unneccesary step in the pipeline if the proper preparation has been made.

(3)
Prepare physical mock-ups of any props used in the sequence (e.g. fishing rod)

(4)
Import the appropriate voice files from the voice recording session.

(5)
Rehearse the performance in the space

(6)
Calibrate the performer in the performance space. This included adding preliminary
joint damping, feet stabilization and shoulder, elbow and hand reaches.

(7)
Load the first character in the sequence and link it to the performer.

(8)
Record takes for the shots or sequences containing the character.

(9)
Load the second and third characters and perform the same, if necessary.

(10)
Export the motion-data in a file containing only the motion data and the character’s skeleton (i.e. remove devices, meshes and “actors’ from the file). At this stage we would also optimize the damping and stabilization settings on a take-by-take basis.