Prior Distribution Design for Music Bleeding-Sound Reduction Based on Nonnegative Matrix Factorization

TL;DR

This paper introduces a novel phase-insensitive blind source separation method using nonnegative matrix factorization with a gamma-distribution prior to effectively reduce bleeding sounds in music recordings.

Contribution

It proposes a new prior-based NMF approach for bleeding-sound reduction that does not rely on phase information, unlike traditional methods.

Findings

Outperforms existing BSS methods in music bleeding-sound reduction

Effective in scenarios where microphones are spatially separated

Utilizes a gamma-distribution prior for leakage levels

Abstract

When we place microphones close to a sound source near other sources in audio recording, the obtained audio signal includes undesired sound from the other sources, which is often called cross-talk or bleeding sound. For many audio applications including onstage sound reinforcement and sound editing after a live performance, it is important to reduce the bleeding sound in each recorded signal. However, since microphones are spatially apart from each other in this situation, typical phase-aware blind source separation (BSS) methods cannot be used. We propose a phase-insensitive method for blind bleeding-sound reduction. This method is based on time-channel nonnegative matrix factorization, which is a BSS method using only amplitude spectrograms. With the proposed method, we introduce the gamma-distribution-based prior for leakage levels of bleeding sounds. Its optimization can be interpreted as maximum a posteriori estimation. The experimental results of music bleeding-sound reduction indicate that the proposed method is more effective for bleeding-sound reduction of music signals compared with other BSS methods.