Interference Reduction in Music Recordings Combining Kernel Additive Modelling and Non-Negative Matrix Factorization

TL;DR

This paper introduces a novel method combining Kernel Additive Modelling and Non-Negative Matrix Factorization to effectively reduce non-stationary interference in music recordings, enhancing separation quality.

Contribution

The paper extends Kernel Additive Modelling with semi-supervised NMF, improving interference localization and kernel frame selection by incorporating musical structure and NMF estimates.

Findings

Improved separation quality over state-of-the-art methods

Effective localization of interference based on NMF activity

Enhanced kernel selection using NMF-based estimates

Abstract

In live and studio recordings unexpected sound events often lead to interferences in the signal. For non-stationary interferences, sound source separation techniques can be used to reduce the interference level in the recording. In this context, we present a novel approach combining the strengths of two algorithmic families: NMF and KAM. The recent KAM approach applies robust statistics on frames selected by a source-specific kernel to perform source separation. Based on semi-supervised NMF, we extend this approach in two ways. First, we locate the interference in the recording based on detected NMF activity. Second, we improve the kernel-based frame selection by incorporating an NMF-based estimate of the clean music signal. Further, we introduce a temporal context in the kernel, taking some musical structure into account. Our experiments show improved separation quality for our proposed method over a state-of-the-art approach for interference reduction.