Generalized Minimal Distortion Principle for Blind Source Separation

TL;DR

This paper introduces a generalized approach to the minimal distortion principle in blind source separation, utilizing mixed-norm models and an efficient optimization algorithm to improve separation quality with minimal computational cost.

Contribution

It proposes a novel mixed-norm model and an iterative reweighted least-squares algorithm for enhanced source separation performance.

Findings

Improved separation by up to 2 dB

No increase in distortion

Low computational cost

Abstract

We revisit the source image estimation problem from blind source separation (BSS). We generalize the traditional minimum distortion principle to maximum likelihood estimation with a model for the residual spectrograms. Because residual spectrograms typically contain other sources, we propose to use a mixed-norm model that lets us finely tune sparsity in time and frequency. We propose to carry out the minimization of the mixed-norm via majorization-minimization optimization, leading to an iteratively reweighted least-squares algorithm. The algorithm balances well efficiency and ease of implementation. We assess the performance of the proposed method as applied to two well-known determined BSS and one joint BSS-dereverberation algorithms. We find out that it is possible to tune the parameters to improve separation by up to 2 dB, with no increase in distortion, and at little computational cost. The method thus provides a cheap and easy way to boost the performance of blind source separation.