Source Separation Evaluation Research Articles

Two‐stage blind audio source counting and separation of stereo instantaneous mixtures using Bayesian tensor factorisation

In this paper, the authors address the tasks of audio source counting and separation for two-channel instantaneous mixtures. This goal is achieved in two steps. First, a novel scheme is proposed for estimating the number of sources and the corresponding channel intensity difference (CID) values. For this purpose, an angular spectrum is evaluated as a function of the ratio of the magnitude spectrogram of the two channels and the peak locations of that spectrum are obtained. In the second stage, a new approach is developed for extracting the individual source signals exploiting a Bayesian non-parametric modelling. The mean field variational Bayesian approach is applied for inferring the unknown parameters. Classification is then performed on the inferred active CID values to obtain the individual source magnitude spectrograms. This way, the number of spectral components used for modelling each source is found automatically from the data. The Bayesian approach is compared with the standard Kullback-Leibler non-negative tensor factorisation method to illustrate the effectiveness of Bayesian modelling. The performance of the source separation is measured by obtaining the existing metrics for multichannel blind source separation evaluation. The experiments are performed on instantaneous mixtures from the dev2 database.

Active-Set Newton Algorithm for Overcomplete Non-Negative Representations of Audio

This paper proposes a computationally efficient algorithm for estimating the non-negative weights of linear combinations of the atoms of large-scale audio dictionaries, so that the generalized Kullback-Leibler divergence between an audio observation and the model is minimized. This linear model has been found useful in many audio signal processing tasks, but the existing algorithms are computationally slow when a large number of atoms is used. The proposed algorithm is based on iteratively updating a set of active atoms, with the weights updated using the Newton method and the step size estimated such that the weights remain non-negative. Algorithm convergence evaluations on representing audio spectra that are mixtures of two speakers show that with all the tested dictionary sizes the proposed method reaches a much lower value of the divergence than can be obtained by conventional algorithms, and is up to 8 times faster. A source separation evaluation revealed that when using large dictionaries, the proposed method produces a better separation quality in less time.

Improved Convolutive and Under-Determined Blind Audio Source Separation with MRF Smoothing

Convolutive and under-determined blind audio source separation from noisy recordings is a challenging problem. Several computational strategies have been proposed to address this problem. This study is concerned with several modifications to the expectation-minimization-based algorithm, which iteratively estimates the mixing and source parameters. This strategy assumes that any entry in each source spectrogram is modeled using superimposed Gaussian components, which are mutually and individually independent across frequency and time bins. In our approach, we resolve this issue by considering a locally smooth temporal and frequency structure in the power source spectrograms. Local smoothness is enforced by incorporating a Gibbs prior in the complete data likelihood function, which models the interactions between neighboring spectrogram bins using a Markov random field. Simulations using audio files derived from stereo audio source separation evaluation campaign 2008 demonstrate high efficiency with the proposed improvement.

Methods for stereo music source separation

Musical source separation from stereo signals has been growing in popularity recently. Algorithms (including the author’s) and the results of a recent stereo audio source separation evaluation campaign (SASSEC) will be presented. In general, such approaches involve extracting mixing parameters from time‐frequency (T‐F) analyses of both channels. Each corresponding T‐F point of the two channels is compared and interchannel intensity differences (IID) and (in the case of stereo mixing), interchannel phase differences (IPD) are computed. A histogram of signal power in the IID and IPD space is constructed, with sources showing strong peaks at the locations of their mixing parameters. Masks are then created for each potential source, corresponding to T‐F points that share common mixing parameters. These masks are applied to the T‐F analysis and resynthesized as probable sources. The underlying assumption of such systems is that of W‐disjoint orthogonality, that is, T‐F points of the sources do not overlap. In music, however, this is rarely the case, as common musical intervals often contain overalapping harmonic ‘‘collisions.’’ The use of an exponentially damped sinusoidal model, that in limited cases has the ability to resolve very closely spaced, beating, sinusoids is explored as a means of separating harmonic collisions.