Temporal and time-frequency correlation-based blind source separation methods. Part I: Determined and underdetermined linear instantaneous mixtures

Abstract

In this paper, we propose two versions of a correlation-based blind source separation (BSS) method. Whereas its basic version operates in the time domain, its extended form is based on the timefrequency (TF) representations of the observed signals and thus applies to much more general conditions. The latter approach consists in identifying the columns of the (permuted scaled) mixing matrix in TF areas where this method detects that a single source occurs. Both the detection and identification stages of this approach use local correlation parameters of the TF transforms of the observed signals. This BSS method, called TIFCORR (for TImeFrequency CORRelation-based BSS), is shown to yield very accurate separation for linear instantaneous mixtures of real speech signals (output SNR's are above 60 dB).

Multi-source TDOA estimation in reverberant audio using angular spectra and clustering

“…We design and conduct a largescale evaluation of angular spectrum-based and clustering-based methods on 1482 different configurations and investigate the use of the former for the initialization of the latter. In addition, we introduce and evaluate five new TDOA estimation methods inspired from signal-to-noise ratio (SNR) weighting or probabilistic modeling techniques that have been successful for anechoic TDOA estimation [19,20,21], histogram-based reverberant TDOA estimation [10] or audio source separation [22,23], but have not yet been explored for angular spectrum-based or clustering-based reverberant TDOA estimation. The proposed methods account for the presence of diffuse noise or interfering sources in each time-frequency bin and rely prioritarily on the time-frequency bins resulting from the direct sound of a single source.…”

Real-Time Multiple Sound Source Localization and Counting Using a Circular Microphone Array

“…Our system performs DOA estimation and source counting assuming there is always at least one active source. This assumption is only needed for theoretical reasons and can be removed in practice, as shown in [33] for example. Additionally, any recent voice activity detection (VAD) algorithm could be used as a prior block to our system.…”

Contribution of statistical tests to sparseness-based blind source separation

“…The elements of this set can be assumed to be isolated time-frequency points as in degenerate unmixing estimation technique (DUET) [15,26] or to form a time-frequency box as in time-frequency ratio of mixtures (TIFROM) [16] and time-frequency CORRelation (TIFCORR) [27]. Assumption 1 is often reasonable thanks to the sparseness of the time-frequency representation of the sources, especially when this number of sources is moderate.…”

“…In TIFCORR [27], the mixing matrix estimation is similar by selecting the empirical covariance coefficients above a certain threshold chosen manually. http://asp.eurasipjournals.com/content/2012/1/169…”

“…In Section "Simulation results", we apply the statistical tests of Section "Statistical tests for sparseness-based UBSS" to several standard UBSS methods [15,16,18,26,27] in the instantaneous mixture case. We thus show that our statistical algorithms reduce the number of empirical parameters and improve the overall performance of the UBSS methods under consideration.…”