Analysis (very rough)

Useful tools for analysis are:

  • Audacity - a graphical tool for audio
  • SoX - command line tool for audio conversion.

Massive hole here.

This page is clearly lacking. The absense of a good open-source, versatile, trainable, classification system for the audio produced by the Solo (and any other field recorders), is stark. This must be solved.

1) open source recogniser / classifier

2) competition / evaluations to appraise and compare algorithms, motivate researchers and groups, and focus attention on what works and what doesn’t. (This requires an freely licensed database of recordings, which isn’t as hard as finding one that everyone is interested in using).

3) easily configurable, trainable systems for conservationists and researchers, who have no interest in learning about classification or machine learning.

4) available to run on any pc/laptop/mac/windows, also deployable in cloud. Also deployable as a client/server setup for inside “offices” or in a University environment, where there are multiple users.

5) The data is too big to move around, so algorithms must run close to your hard drive.

6) Consideration of ultra and infrasound.