Trials on constructing an accelerometer-based conducting system
MetadataShow full item record
Computer-based conducting systems are augmented music performance direction. Gestural interfaces have been used to provide a natural way to manipulate the system. These systems employed various sensors to capture gesture data. Collected sample data was then analyzed and used to identify gestures. This presentation provides some trials on construcing a conducting system based on an accelerometer sensor. The movement of the right arm of a conductor is traced by an accelerometer sensor. The sample data is received via bluetooth and saved into a text file. 2-dimensional data are extracted from 3-dimensional data. Then the feature sequence is fed into a model to calculate two parameters of that model. This model can be used to classify a conducting gesture. Our system supports three beat-patterns: 2-beat conducting gesture, 3-beat conducting gesture and 4-beat conducting gesture. Each beat pattern corresponds to a model. The system is intended to aid the learning and practice of conducting gestures which is only supported by a few systems before. The system implementation is at a preliminary stage. Many tasks are able to be done in future. Different models can be used to represent beat patterns. Camera, instead of or combined with the accelerometer sensor, will be used to collect sample data. This presentation introduces our trials on constructing an accelerometer-based conducting system to help the learning and practice of conudcting gestures.