Combining accelerometer and video camera: Reconstruction of bow velocity profiles

Erwin Schoonderwaldt (1, 2) , Nicolas H. Rasamimanana (1, 2) , Frédéric Bevilacqua (1, 2)

1 Equipe Interactions musicales temps-réel
2 Equipe Autre (hors R&D)

A cost-effective method was developed for the estimation of the bow velocity in violin playing, using an accelerometer on the bow in combination with point tracking using a standard video camera. The video data are used to detect the moments of bow direction changes. This information is used for piece-wise integration of the accelerometer signal, resulting in a drift-free reconstructed velocity signal with a high temporal resolution. The method was evaluated using a 3D motion capturing system, providing a reliable reference of the actual bow velocity. The method showed good results when the accelerometer and video stream are synchronized. Additional latency and jitter of the camera stream can importantly decrease the performance of the method, depending on the bow stroke type.

Keywords: violin; Bowing gestures; bowed string; bow velocity; accelerometer; video tracking

Keywords: Informatique musicale

Domains: Sound [cs.SD]; Human-Computer Interaction [cs.HC]; Musicology and performing arts; Computer Science [cs]; Signal and Image Processing; Machine Learning [cs.LG]; Artificial Intelligence [cs.AI]; Computer Aided Engineering; Multimedia [cs.MM]; Computer Vision and Pattern Recognition [cs.CV]; Other [cs.OH]; Signal and Image processing

Erwin Schoonderwaldt, Nicolas H. Rasamimanana, Frédéric Bevilacqua. Combining accelerometer and video camera: Reconstruction of bow velocity profiles. 6th International Conference on New Interfaces for Musical Expression, 2006, NA, France. pp.1-1. ⟨hal-01161363⟩