Title :

A frugal approach to music source separation

Author(s) :

Pierson Lancaster, Emery [Auteur]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]
Souviraà-Labastie, Nathan [Auteur]
A-Volute [Roubaix]

English keyword(s) :

Index Terms-Audio separation
audio deep learning
audio database
music source separation
benchmark

HAL domain(s) :

Informatique [cs]

English abstract : [en]

During the past years, deep learning brought a big step in performance of music source separation algorithms. A lot has been done on the architecture optimisation, but training data remains an important bias for model ...
Show more >During the past years, deep learning brought a big step in performance of music source separation algorithms. A lot has been done on the architecture optimisation, but training data remains an important bias for model comparison. In this work, we choose to work with the frugal and well-known original TasNet neural network and to focus on simple methods to exploit a relatively important dataset. Our results on the MUSDB test set outperform all previous state of the art approaches with extra data on the following source categories: vocals, accompaniment, drums, bass and in average. We believe that our results on how to shape a training set can apply to any type of architecture.Show less >

Language :

Anglais

Collections :

• Centre de Recherche en Informatique, Signal et Automatique de Lille (CRIStAL) - UMR 9189

Source :

Harvested from HAL