Abstract
Audio source separation is the process of isolating individual sonic elements from a mixture or auditory scene. We present the Northwestern University Source Separation Library, or nussl for short. nussl (pronounced ‘nuzzle’) is an open-source, object-oriented audio source separation library implemented in Python. nussl provides implementations for many existing source separation algorithms and a platform for creating the next generation of source separation algorithms. By nature of its design, nussl easily allows new algorithms to be benchmarked against existing algorithms on established data sets and facilitates development of new variations on algorithms. Here, we present the design methodologies in nussl, two experiments using it, and use nussl to showcase benchmarks for some algorithms contained within.
| Original language | English (US) |
|---|---|
| Title of host publication | Proceedings of the 19th International Society for Music Information Retrieval Conference, ISMIR 2018 |
| Editors | Emilia Gomez, Xiao Hu, Eric Humphrey, Emmanouil Benetos |
| Publisher | International Society for Music Information Retrieval |
| Pages | 297-305 |
| Number of pages | 9 |
| ISBN (Electronic) | 9782954035123 |
| State | Published - Jan 1 2018 |
| Event | 19th International Society for Music Information Retrieval Conference, ISMIR 2018 - Paris, France Duration: Sep 23 2018 → Sep 27 2018 |
Publication series
| Name | Proceedings of the 19th International Society for Music Information Retrieval Conference, ISMIR 2018 |
|---|
Conference
| Conference | 19th International Society for Music Information Retrieval Conference, ISMIR 2018 |
|---|---|
| Country/Territory | France |
| City | Paris |
| Period | 9/23/18 → 9/27/18 |
Funding
We have presented the Northwestern University Source Separation Library (nussl), an open-source, object-oriented audio source separation library implemented in Python. nussl implements many popular source separation algorithms, and a low barrier API for end-users and developers alike. We have demonstrated its design framework, including its ability to interface with common data sets and evaluation metrics. We also showcased two novel experiments using the API and a set of benchmarks. This project is actively seeking submissions from eager researchers and avid open source developers. Readers can find more information at interactiveaudiolab. github.io/nussl. This work was supported by USA National Science Foundation Award 1420971.
ASJC Scopus subject areas
- Music
- Information Systems