Musculoskeletal models can be used to distinguish too short or too slow muscles in Cerebral Palsy patients. Therefor muscle-tendon geometry needs to be modeled accurately in order to predict the need and effectiveness of muscle-tendon lengthening. This geometry is strongly related to the location and orientation of the joint centres. In clinic joint centres are determined with the used of anatomical marker positions combined with regression equations, or using functional calibration methods. The method of choice is known to strongly influence join centre locations. Little is known about the functional relevance of these difference. In this study it was investigated to what extent marker-based versus functional calibration methods affect muscle-tendon length and velocity in modeling gait.