Jeff Snyder

The Birl: Adventures in the Development of an Electronic Wind Instrument

This article reflects on the markedly distinct development stages of an electronic wind instrument called the Birl. Stemming from an early idea for an electro-mechanical oscillator inspired by the sounds of pen plotters, the Birl was formed through the connection of that oscillator prototype to a rough wind instrument body. Originally intended to fulfill the role of the wind section in an ensemble of instruments built for the author’s doctoral dissertation composition, the instrument took on a new life after the completion of the piece. The development of a “cello-like” resonator body and refinements to the electro-mechanical aspects had brought the instrument to a performable state, but several limitations suggested further development. A desire to make the instrument more conducive to exploratory improvisation pushed the Birl in new directions, toward open-holed fingering systems and embouchure sensors with neural net mapping structures and physical models of dynamically configurable toneholes, resulting in an instrument that bore little resemblance to the original electro-mechanical concept. The author discusses the design challenges that arose as the instrument evolved, the solutions that were found along the way, and the ways in which user feedback informed the design as the needs of the instrument changed.

Learning to identify container contents through tactile vibration signatures

We examine using a simple contact sensor coupled with standard machine learning algorithms to classify and count objects shaken in a container. The contact sensor measures the resulting vibrations, and these signatures are used to learn a classifier that maps vibration signatures to known object categories. A linear support vector machine trained on labeled vibration signatures achieves a mean binary classification accuracy of 99% over 66 pairs of objects and a mean multi-class classification accuracy of 94% over 12 classes. It is also shown that useful tasks such as approximate counting of objects over the range 1 to 10 is possible. We see potential applications of these ideas in service robots engaged in cleanup and inventory control in labs, workshops, stores, warehouses and homes.