About Birdy Music
Birdy Music is an audio implementation of real-time bird sound detection and classification, built upon the foundational work of the BirdNET project.
BirdNET is a convolutional neural network based bird sound classification system trained on annotated avian audio samples to identify species from field recordings with high accuracy. It was developed by the Cornell Lab of Ornithology and Chemnitz University of Technology. You can read more about it here https://ceur-ws.org/Vol-1866/paper_143.pdf
Audio synthesis and bussing is facilitatied with the wonderful Tone.js library.
Neural frequency coupling and spatial arrangment is implemented using recent research by Rosso and the IPEM Institute for Systematic Musicology and Ghent University. Read all about it here https://advanced.onlinelibrary.wiley.com/doi/epdf/10.1002/advs.202413195
Field recordings are made on a $10 dollar omnidirectional mic and Raspberry Pi 4b and pushed to a Ruby on Rails app hosted on a VPS where tagging and classification of traits are made to facilatate algorhythmic musical generation to fill your ear holes with warm hugs and cool kisses.