04 Mocap Delivery Examples

Lesson Summary

  • Typical Mocap Delivery Format:
    • Mocap is usually delivered on a simple FK (forward kinematic) skeleton with no controllers or driven joints.
    • The data will have been solved (converted) from translational or rotational data from optical (Vicon, OptiTrack) or inertial (MVN, Rokoko) mocap systems.
  • Data Should Be Clean:
    • Data should be cleaned to fix any errors introduced during capture.
    • Common errors include noise or jitter caused by marker tracking issues, and swaps and flips where markers are incorrectly identified causing joints to be wrongly oriented for some or all of the take.
  • Full vs Trimmed Mocap Takes:
    • Mocap takes may contain multiple attempts of a single take.
    • Using timecode to identify specific ranges of mocap data for processing helps reduce costs.
  • Diagonal Motion Capture:
    • Performing actions diagonally in the capture volume allows fitting longer actions or more steps into a smaller space.
  • Understanding Capture vs. Scene Space:
    • Motion capture data is often delivered in capture space (where it was recorded).
    • The data may need realigning to scene space for technical reason or to align with an environment.
  • Capture Frame Rate vs Delivery Frame Rate:
    • Motion capture is normally captured at a higher frame rate than the final delivery frame rate e.g. 120 vs 24.
    • Editing at the delivery frame rate ensures accurate adjustments without unnecessary work on unseen frames.


performance_capture_pipeline.pdf
Complete and Continue  
Discussion

2 comments