This paper explores the potential of prioritized Inverse Kinematics for full body motion capture. We are especially interested in compensating for the use of a minimal set of sensors by proposing pertinent constraints associated with a priority level. We evaluate the believability issues and the performance requirements arising through a case study exploiting recorded motions of professional clarinet performers. We conclude by presenting a framework in which end users can easily configure the IK engine for 3D real-time interaction.