In 2016 I was contacted by Dan Stowell, a machine learning and bioacoustic researcher at QMUL. Dan had created a system for automatically transcribing birdsong using machine-learning algorithms. His software created a stream of frequencies, amplitudes and timepoints which really lent themselves to musical adaptation.

I created a Max patch to convert his data into a format that could be interpreted automatically by the Ealing Feeder, my robotic carillon. We used it to play Dan’s automatic transcriptions of a dawn chorus at Stave Hill Ecology Park on Interntional Dawn Chorus Day.

It’s very interesting to compare this automatic transcription method to the eighteenth century interpretations of birdsong in the Bird Fancyer’s Delight.

Thanks to Karina Townsend for this video of the carillon playing outdoors at Stave Hill.