Welcome to FishSound Finder!

FishSound Finder is an open source python software to automatically detect and classify fish sounds in passive acoustic recordings. It is based on the python libraries ecosound and ketos. FishSound Finder was trained to detect typical low frequency (< 1kHz) isolated grunts and knocks produced by fish. Future versions will also include the capability to detect sounds from specific fish species (e.g. haddock). It was originally developed to analyze fish sounds found in coastal British Columbia (Canada) but proved to also work well in other parts of the world (e.g. Florida, USA). Detection outputs from FishSound Finder are compatible with the popular bioacoustics software Raven.

Content:

• FishSound Finder
  • Why FishSound Finder?
  • What can I use it for?
  • Does it work in all environments?
  • How does it work?
  • Who can use it?
  • Requirements
  • Contributors
• Installation
  • Short version
  • Longer version (python beginners)
• Usage
  • Required arguments
  • Optional arguments
  • Example
• Tutorial
  • Setting up the project structure
  • Downloading the model file
  • Downloading audio data
  • Creating a deployment file
  • Running FishSound Finder
  • Reviewing the processing logs
  • Analyzing the detection results (documentation in progress)
• Contributing
  • How can you contribute ?
  • What is the best way to reach out to us?
• Credits

Status

FishSound Finder has been tested on a number of different environments and is considered stable. We are keen to keep improving FishSound Finder and encourage users to get in touch with us to provide feedback and report issues/bugs.

Contributors

Lead developper:

• Xavier Mouy

Collaborators:

• Dana Haggarty

• Francis Juanes

• Sarah Dudas

• Sofie Van Parijs

• Stephanie Archer

• Darienne Lancaster

• Philina English

• Bill Halliday

Analysts :

• Aislyn Adams

• Cierra Hart

• Courtney Evans

• Emie Woodburn

• Erik Archer

License

FishSound Finder is licensed under the open source BSD-3-Clause License.