This repository is concerned with developing novel ways to computationally work with musical, especially percussive, sound onsets.
Currently several items are developed concurrently:
- Percussive Onset Sound Dataset specification (POSD)
- an open dataset specification to methodically work with audio onset data
- see notebooks/dataset_spec_draft.org
- Multilateration algorithms to locate where sounds were generated based on
multiple sensors
- e.g. when striking a drumhead, use several sensors to get precise coordinates of where the strike landed
- Onset fingerprinting/classification algorithms
- given a percussive onset, classify it as being a specific sound, under real-time constraints
- used for example to electronically augment live playing