Open-Source Tools & Data for Music Source Separation

  • Open Source Tools & Data for Music Source Separation

Introduction

  • Tutorial Structure and Setup
  • What is Source Separation?
  • Map of Open-Source Source Separation Projects

Basics of Source Separation

  • Representing Audio
  • TF Representations and Masking
  • Phase
  • Evaluation

First Steps

  • Introduction to nussl
  • Putting it All Together: Repetition
  • Build Your Own HPSS

Deep Learning Approaches

  • Introduction
  • Building Blocks
  • Architectures

Data

  • Introduction
  • Datasets
  • The MUSDB18 dataset
  • Generating mixtures with Scaper
  • Recap

Training

  • Introduction
  • Gradient descent
  • Coding up model architectures
  • Putting it all together

Conclusions

  • Applications
  • Concluding Remarks

Resources

  • References
  • Additional Reading
  • Cite this Book
  • Acknowledgements
  • About the Authors
Powered by Jupyter Book
Contents
  • Source Separation
  • Learning about Music Information Retrieval (MIR)

Additional Reading¶

Source Separation¶

Additional resources for learning more about source separation:

  • SigSep Website

  • DCASE Task 4 - Environmental Sound Separation

  • Emmanuel Vincent’s website

Learning about Music Information Retrieval (MIR)¶

Other resources useful for further exploration of MIR topics:

  • Music Information Retrieval website

  • Prof. Bryan Pardo’s class on Machine Perception of Music

  • Prof. Alexander Lerch’s Audio Content Analysis website

  • Coursera Course on Audio Signal Procesing for Musical Applications

  • Prof. Julius O. Smith III’s Online DSP Textbooks

  • Fabian-Robert Stöter’s Awesome Python Scientific Audio Github list

References Cite this Book

By Ethan Manilow, Prem Seetharaman, Justin Salamon
© Copyright 2020.