Kinematics of Guitarists

Abstract

Numerous studies have shown that synergies from kinematic data are reusable for reconstructing original movements, even building new ones. The concept of synergy has drawn much attention in modern fields, especially in robotics. Surprisingly, time-varying synergies extraction can also apply to instrument player dexterity study. This data-driven study investigates any subtle kinematic differences between novice and expert guitar players' finger movements on both repetitive chromatic scale and advanced musical scale playing. Principal Component Analysis was adopted to extract synergies from all subject simultaneously. Support Vector Machine Recursive Feature Elimination (SVM-RFE) was further applied on the weightings of extracted common patterns, which successfully achieved an accurate classification of the finger movements from novice and expert player subjects. It was found that fundamental finger moving patterns of novices and experts are similar but, according to signals reconstructed using synergies selected by the SVM-RFE algorithm, novices generally have more redundant movements in irrelevant fingers and the pinkie finger may show the largest range of distinction in all four long finger reconstructions. The revelations of this study may provide a scientific and efficient direction of teaching and learning the guitar, even contribute to producing mechanical systems that can represent dexterous human finger movement using synergistic approach.