Weakly Informed Audio Source Separation

Abstract : Prior information about the target source can improve audio source separation quality but is usually not available with the necessary level of audio alignment. This has limited its usability in the past. We propose a separation model that can nevertheless exploit such weak information for the separation task while aligning it on the mixture as a byproduct using an attention mechanism. We demonstrate the capabilities of the model on a singing voice separation task exploiting artificial side information with different levels of expres-siveness. Moreover, we highlight an issue with the common separation quality assessment procedure regarding parts where targets or predictions are silent and refine a previous contribution for a more complete evaluation.
Document type :
Preprints, Working Papers, ...
Complete list of metadatas

Cited literature [29 references]  Display  Hide  Download

https://hal.telecom-paristech.fr/hal-02332689
Contributor : Kilian Schulze-Forster <>
Submitted on : Friday, October 25, 2019 - 7:11:05 AM
Last modification on : Wednesday, October 30, 2019 - 1:37:32 AM

File

WASPAA_paper_2019_HAL.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02332689, version 1

Citation

Kilian Schulze-Forster, Clément Doire, Gaël Richard, Roland Badeau. Weakly Informed Audio Source Separation. 2019. ⟨hal-02332689⟩

Share

Metrics

Record views

27

Files downloads

31