Accelerometers as an Instrument
summary
Hand-worn devices transmit three-axis accelerometer data over Wi‑Fi from a microcontroller to a computer and drive gesture-to-sound mapping. Sound is rendered in third-order ambisonics decoded for eight speakers plus a sub (nine channels). Corpus-based synthesis navigates 2D sound maps by hand tilt. A recursive neural network (Magenta) is used to make the piece adapt but is trained offline and not reliably live; the hardware-and-network setup successfully performs the prepared piece tonight.
AI-written summary · about
chapters (9)
- 0:00Welcome
- 1:25Speaker intro and demo
- 2:53Computer music history
- 4:57Accelerometer instruments evolution
- 6:23Spatial sound setup
- 7:47Accelerometer tech details
- 9:12Sound corpus and ML
- 11:28Performance demo
- 19:27Q&A
related talks
33mUsing mobile technology in live musical perfomance
41mCreating Intelligent Music Instruments with Machine Learning
08mNo merge conflicts
39mSonic Earth - An Algorithmic Performance Approach to Real-Time Environmental Sonification
28mStrike a Pose - Gesture recognition in JavaScript w/ ML & Arduino
11mPaper Synthesizer! Music created with Augmented Reality
18mResponsive Space. Liveness through spatial distribution of sound and image
31mEarthbeat: The Art of Creating your own Senses