[ylwm_vimeo]56952053[/ylwm_vimeo]

Our research has already started. We have tried two different softwares to work with the Microsoft Kinect, Ethno Tekh’s Ethno Tracker and FAAST. While Talía played the violin, we observed the movements and parts of the body that each skeleton tracker detects. Elbow, wrist and hand were to be analysed. Having tested different bow strokes from different perspectives, we noticed that the Ethno Tracker collected the signal better when the violin player was in front of the machine. Here we have the results of the different bowing techniques:

• Spiccato: elbow and speed of the right hand are detected, but without a lot of variation in the position.
• Staccato: speed of the right hand and elbow position are detected.
• Detaché: speed of the right hand and elbow position are detected.
• String crossing: speed and position of the elbow and the right hand are detected.
• Martelé: speed and position of the elbow and the right hand are detected.

The results were quite satisfactory. We have already seen what is possible to analyse with Kinect. The next step is trying to train the machine to recognize the bow strokes for itself. This is step is not compelling yet to draw the graphics in the interface.

Let’s keep up the good work!