\ \NoIze/ /

Welcome to the docs!

NoIze is a Python package for accessible smart noise filter development. Interested? Here is a demo! NoIze also offers stand-alone sound filtering as well as sound classifier tools. It all started as a Prototype Fund project from the 5th round: “Letting Machines learn–Technologies for our futures”. Now it is open source and available as a tool for learning or development.

This package is a prototype for the purpose of empowering listers’ experience of their sound environment. The foundation has been laid for users to collect sounds from their surroundings and implement noise filters uniquely designed not only according to their own surroundings, but also to their own perceptions of noise.

The structure of the convolutional neural network built in NoIze, used as a noise classifier, was inspired by the researchers A. Sehgal and N. Kehtarnavaz in their 2017 paper. The filtering functionality stems from another paper from the same university, University of Texas at Dallas, with an additional author: A. Bhattacharya.


Peggy Sylopp



Aislyn Rose