Tools of soundscape information retrieval

Author: Tzu-Hao Harry Lin ([email protected])

This is a python version of Soundscape_Viewer (originally MATLAB-based tool), which aims to facilitate the visualization, source separation, and event identification of long-duration field recordings.

In order to facilitate the cloud computation, several functions were added to assist the data retrieval from cloud-based platforms (e.g., Google Drive). Now you can access the audio data archived on pumilio-based soundscape archive (e.g., Asian Soundscape) and visualize a large amount of audio data by using the long-term spectrograms. You can also employ the source separation to imporve the acoustic analysis result. Finally, apply a clustering tool to investigate the acoustic phenology.

Future works:

1. Support interactive visualization of clustering result (under development).
2. More source separation and clustering functions will be added (scheduled at 2020 winter).

Demo (in Google Colab, recent update: 2019/9/27): https://colab.research.google.com/drive/1Bod1SW8YtVgFWPO4w8WFGORmku-MPCL-