This is a user friendly guide to analysing wav. audio files for bird and bat species diversity in Jupyter notebooks.
With the rise in easily accessable recording devices, I thought it would be great to have an easy option available for less-code minded people to analyse their files for species. If you are interested in monitoring for birds and bats yourself, I would reccomend using an AudioMoth https://www.openacousticdevices.info/audiomoth.
After you have installed all the necessary packages and saved the analyzer.py files in your local files, you can use the command line implementation to analyse your files in one line of code.
I did not create the models, only wrote the code for implementation.
The Jupyter notebooks implementation will analyse your folder of audio clips and produce a CSV file of the Bird species with the "recording_id", "common_name", "scientific_name", "start_time", "end_time" and "confidence" as columns.
CLI implementation:
!python bird_analyzer_cli.py \ --min-conf 0.8 \ --lat 55.4 --lon 3.4 \ input_directory \ output_csv.csv
All code relating to the model and downloading the required packages can be found here https://github.com/kahst/BirdNET-Analyzer/blob/main/README.md.
The Jupyter notebooks implementation will analyse your folder of audio clips and produce a CSV file of the Bat species with the "recording_id", "start_time", "end_time", "low_freq", "high_freq", "class", "class_prob" and "det_prob" as columns.
CLI implementation:
!python bat_analyzer_cli.py \ --detection-threshold 0.3 \ --time-expansion-factor 1 \ --max-duration 3 \ input_folder \ output_file.csv
All code relating to the model and downloading the required packages can be found here https://github.com/macaodha/batdetect2/blob/main/batdetect2_notebook.ipynb.
This is a great package that unlockes the associated meta data with each audio moth file. https://github.com/mbsantiago/metamoth