A soundscape is an acoustic landscape formed by sounds of biological, environmental, and anthropogenic sources. With a multitude of acoustic information associated with animal behaviors, environments, and human activities contained in soundscapes, researchers can predict biodiversity change purely by passive listening. However, acoustic analysis remains laborious due to noise interference and the variability of animal vocalizations.
Dr. Tzu-Hao Lin's research team (Biodiversity Research Center, Academia Sinica) and Dr. Shih-Ching Yen (Center for General Education, National Tsing Hua University) developed soundscape_IR, an open-source Python toolbox dedicated to soundscape information retrieval, for facilitating the assessment of acoustic diversity. The results showed that soundscape_IR streamlined the process of detecting sika deer calls from tropical forest recordings and revealed their repertoire structure. Furthermore, soundscape_IR automated the investigation of biotic and abiotic sounds, enabling an efficient assessment of community divergence between marine habitats. In addition to audio source separation, soundscape_IR also supports functions for visualizing soundscape dynamics, automatic feature extraction, and an easy-to-use batch processing module for analyzing long-duration recordings.
This study made the cover image of the 2022 November issue of Methods in Ecology and Evolution.
Sun, Y.-J., Yen, S.-C., Lin, T.-H. (2022) soundscape_IR: A source separation toolbox for exploring acoustic diversity in soundscapes. Methods in Ecology and Evolution, 13(11): 2347-2355. https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.13960
Codes of soundscape_IR:
Technical documentation and examples: