Skip to content

polimi-ispl/vbe-demixing

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 

Repository files navigation

Virtual Bass Enhancement Via Music Demixing

Abstract: Virtual Bass Enhancement (VBE) refers to a class of digital signal processing algorithms that aim at enhancing the perception of low frequencies in music applications. Such algorithms are particularly valued for improving the performance of small-size transducers often found in consumer electronics and typically exploit well-known psychoacoustic effects. Though both time- and frequency-domain techniques have been proposed in the literature, none of them capitalizes on the latest achievements of deep learning as far as music processing is concerned. In this letter, we propose a novel time-domain VBE algorithm that incorporates a deep neural network for music demixing as part of the processing pipeline. This technique is shown to improve the bass perception and reduce inharmonic distortion, i.e., the main issue of existing time-domain VBE algorithms. The results of a perceptual test are then presented, showing that the proposed method is able to outperform state-of-the-art algorithms both in terms of bass enhancement and basic audio quality.

The paper is available at this link.

Credits

R. Giampiccolo, A. I. Mezza, A. Bernardini, and A. Sarti, "Virtual Bass Enhancement via Music Demixing,” IEEE Signal Processing Letters, vol. 30, pp. 908-912, 2023.

@inproceedings{giampiccolo:vbe2023, author = {R. Giampiccolo, and A. I. Mezza, and A. Bernardini, and A. Sarti}, journal = {IEEE Signal Processing Letters}, title = {Virtual Bass Enhancement via Music Demixing}, pages = {908-912}, month = {Jul.}, year = {2023}, doi={10.1109/LSP.2023.3296877}}

Releases

No releases published

Packages

No packages published