Building Guitar Strum Models for an Interactive Air Guitar Prototype

Towards Developing a Virtual Guitar Instructor through Biometrics Informed Human-Computer Interaction

“…Other composers are using wearable interfaces, such as the Myo armbands, to embody and sonify (create sound from data) physiological information both within the physical [31] and digital spaces [7] (including mixed reality (MR) [8]). Other authors using Myo armbands to classify musical gesture have reported the potential improtance of using VR to visualise musical performance information [30]. Using wearable sensors to sonify bodily data allows such composers to find new methods for composing music and gesturally control sound more naturally (e.g., such as the intentions behind the much earlier Theremin).…”

“…Therefore, pairing such wearable interfaces with game engines allows us to investigate how musical technique can be communicated intuitively to student musicians (here, guitarists), which is the overarching goal and intended future path for this research enquiry. This work is motivated by similar literature in the field, which investigates the classification of gestural information for musical instrument performance [9,10,30], towards understanding applications for doing so. The problems existing in the domain of musical performance and HCI concern methodologies and best practices for processing novel and more complex biometric information, such as EMG.…”

“…Authors have successfully used deep learning models to classify orientation data, in partnership with professional musicians, but EMG information has seldom been used with deep learning algorithms [10]. Other authors have classified EMG biometrics with general ML algorithms towards creating a guitar which can be played in the 'air', where the problems outlined are the accuracy of the predictions and determining the sample size [30].…”