Virtual Analog Modeling Using Neural ODEs

This page contains audio examples for the "Virtual Analog Modeling Of Distortion Circuits Using Neural Ordinary Differential Equations" paper presented at the 25th International Conference on Digital Audio Effects (DAFx20in22) in Vienna, Austria, September 2022.

You can check out this publication on arxiv or on the conference's website.

Please, check out also the GitHub repository for this publication.

If you have any questions or remarks, please, word them in the comments section at the bottom.

Bibtex Citation

When you cite this work, please use the following bibliographical data:

@InProceedings{Wilczeketal2022,
    author    = {Wilczek, Jan and Wright, Alec and Välimäki, Vesa and Habets, Emanuël},
    booktitle = {Proceedings of the 25th International Conference on Digital Audio Effects (DAFx20in22), Vienna, Austria, September 2020-22},
    title     = {Virtual {A}nalog {M}odeling of {D}istortion {C}ircuits {U}sing {N}eural {O}rdinary {D}ifferential {E}quations},
    year      = {2022}
  }

Authors

• Jan Wilczek
• Alec Wright
• Vesa Välimäki
• Emanuël Habets

Abstract

Recent research in deep learning has shown that neural networks can learn differential equations governing dynamical systems. In this paper, we adapt this concept to Virtual Analog (VA) modeling to learn the ordinary differential equations (ODEs) governing the first-order and the second-order diode clipper. The proposed models achieve performance comparable to state-of-the-art recurrent neural networks (RNNs) albeit using fewer parameters. We show that this approach does not require oversampling and allows to increase the sampling rate after the training has completed, which results in increased accuracy. Using a sophisticated numerical solver allows to increase the accuracy at the cost of slower processing. ODEs learned this way do not require closed forms but are still physically interpretable.

Audio Examples

All audio examples were normalized to -23 LUFS using the pyloudnorm library from Christian Steinmetz.

For example display the trackswitch.js library was used:

Werner, Nils, et al. "trackswitch.js: A Versatile Web-Based Audio Player for Presenting Scientifc Results." 3rd web audio conference, London, UK. 2017.

Table of Contents

1. First-Order Diode Clipper
1. Bass1
2. Bass2
3. Guitar1
4. Guitar2
2. Second-Order Diode Clipper
1. Bass1
2. Bass2
3. Guitar1
4. Guitar2

First-Order Diode Clipper

Bass1

Bass2

Guitar1

Guitar2

Second-Order Diode Clipper

Bass1

Bass2

Guitar1

Guitar2