How can computers learn to recognize birds from sounds? The Cornell Lab of Ornithology and the Chemnitz University of Technology are trying to find an answer to this question. Our research is mainly focused on the detection and classification of avian sounds using machine learning – we want to assist experts and citizen scientist in their work of monitoring and protecting our birds.
This page features some of our public demonstrations, including a live stream demo, a demo for the analysis of audio recordings and an Android app. All demos are based on an artificial neural network we call BirdNET. We are constantly improving the features and performance of our demos – please make sure to check back with us regularly.
We are currently featuring more than 500 of the most common species of North America and Germany. We will add more species and more regions in the near future. Click here for the list of supported species.
Live Stream Demo
The live stream demo processes a live audio stream from a microphone outside the Cornell Lab of Ornithology, located in the Sapsucker Woods sanctuary in Ithaca, New York. This demo features an artificial neural network trained on the 75 most common species of the Sapsucker Woods area. Our system splits the audio stream into segments, converts those segments into spectrograms (visual representations of the audio signal) and passes the spectrograms through a convolutional neural network, all in near-real-time. The web page accumulates the species probabilities of the last five seconds into one prediction. If the probability for one species reaches 15% or higher, you can see a marker indicating an estimated position of the corresponding sound in the scrolling spectrogram of the live stream. This demo is intended for large screens.
Follow this link to view the demo.
Analysis of Audio Recordings
Reliable identification of bird species in recorded audio files would be a transformative tool for researchers, conservation biologists, and birders. This demo provides a web interface for the upload and analysis of audio recordings. Based on an artificial neural network featuring more than 500 of the most common species of North America and Germany, this demo shows the most probable species for every second of the recording. Please note: We need to transfer the audio recordings to our servers in order to process the files. This demo is intended for large screens.
Follow this link to view the demo.
Click here to download a demo recording.
Android App
This demo lets you record a file using the internal microphone of your Android device and an artificial neural network will tell you the most probable bird species present in your recording. We use the native sound recording feature of smartphones and tablets as well as the GPS-service to make predictions based on location and date. Give it a try! Please note: We need to transfer the audio recordings to our servers in order to process the files. Recording quality may vary depending on your device. External microphones will probably increase the recording quality.
Follow this link to download the app.
Note: We consider our app a prototype and by no means a final product. If you encounter any instabilities or have any question regarding the functionality, please let us know. We will add new features in the near future, you will receive all updates automatically.