Time-domain Blind Source Separation Research Articles

A Time-Domain Algorithm for Blind Source Separation of Non-stationary Convolutive Mixtures

In this paper, a time-domain blind source separation algorithm for non-stationary convolutive mixtures is proposed by reprogramming the vectors of convolutive mixture model and generalizing the joint approximate diagonalization method. Firstly the sampling convolutive mixture signals are reseted for matching instantaneous mixture model, then considering non-stationarity of the sources, space whitening and joint block-diagonalization method is exploited to obtain the original signals. This algorithm simplifies the convolutive mixture problem into the instantaneous mixture problem from a new point of view, so it avoids domain transformation and convolution operation, as well as decreases the complexity. Computer simulation verifies its effectiveness and gives the analysis results about the effect on the signal to interference ratio as its parameter changes.

A time domain algorithm for blind separation of convolutive sound mixtures and L1 constrainted minimization of cross correlations

A time domain blind source separation algorithm of convolutive sound mixtures is studied based on a compact partial inversion formula in closed form. An L1-constrained minimization problem is formulated to find demixing filter coefficients for source separation while capturing scaling invariance and sparseness of solutions. The minimization aims to reduce (lagged) cross correlations of the mixture signals, which are modeled stochastically. The problem is non-convex, however it is put in a nonlinear least squares form where the robust and convergent Levenberg-Marquardt iterative method is applicable to compute local minimizers. Efficiency is achieved in recovering lower dimensional demixing filter solutions than the physical ones. Computations on recorded and synthetic mixtures show satisfactory performance, and are compared with other iterative methods.

Partial separation method for solving permutation problem in frequency domain blind source separation of speech signals

This paper addresses the well known permutation problem in frequency domain blind source separation. The proposed method uses correlation between two signals in each DFT bin to solve the permutation problem. One of the signals is partially separated by a time domain blind source separation method and the other is obtained by the frequency domain blind source separation method. Two different ways of configuring the time and frequency domain blocks, i.e., in parallel or cascade, have been studied. The cascaded configuration not only achieves a better separation performance but also reduces the computational cost as compared to the parallel configuration.

A general contrast function based blind source separation method for convolutively mixed independent sources

This paper presents a time-domain blind source separation algorithm based on the criterion of the general contrast function. The algorithm extends the efficient fixed-point algorithm for instantaneous mixing models to extract a source from convolutive mixed sources. A new decorrelation procedure is proposed to help the algorithm separate the rest of the sources one by one. This scalable feature enables us to extract only some of the source signals from their convolutive mixtures. Also, the algorithm does not suffer from the problems of frequency domain based algorithms on arbitrary scaling and permutations in different frequency components. Simulations on synthetic data and real-world recordings were used to verify the extended fixed-point algorithm. Results of the simulations show that the algorithm is capable of significantly suppressing the crosstalk of the extracted source signals.

Blind Source Separation of Convolutive Mixtures of Speech in Frequency Domain

This paper overviews a total solution for frequency-domain blind source separation (BSS) of convolutive mixtures of audio signals, especially speech. Frequency-domain BSS performs independent component analysis (ICA) in each frequency bin, and this is more efficient than time-domain BSS. We describe a sophisticated total solution for frequency-domain BSS, including permutation, scaling, circularity, and complex activation function solutions. Experimental results of 2 × 2, 3 × 3, 4 × 4, 6 × 8, and 2 × 2 (moving sources), (#sources × #microphones) in a room are promising.