Blind separation of multiple speakers in a multipath environment
نویسندگان
چکیده
We relate information theoretic blind learning methods (infomax) and Bussgang blind equalization methods. The multipath extension of blind source separation methods can be seen in the frequency domain using FIR matrix algebra (matrices of nite impulse response lters). Three forms of Bussgang algorithms are given. The blind serial update method of Cardoso and Laheld is related to the infomax objective of Bell and Sejnowski. The application emphasis is on speech separation. We demonstrate the robustness and power of the new techniques by blindly separating speech signals recorded in a multipath environment.
منابع مشابه
A graph theoretical framework for consistent time differences of arrival
The estimation of time difference of arrival (TDOA) plays an important role in acoustic source localization, blind source separation, acoustic sound control, and improved speech communication. However, when multiple sources are active in a reverberant environment, the TDOA estimates are known to suffer from the multipath and multiple source ambiguity, and hence are less reliable. In a previous ...
متن کاملBlind identification of FIR channels carrying multiple finite alphabet signals
The finite alphabet property of digital communication signals, along with oversampling techniques, enables the blind identification and equalization of an unknown FIR channel carrying a superposition of such signals, provided they have the same (known) period. Applied to multi-user wireless communications, the same framework allows the blind separation of multiple finite alphabet signals receiv...
متن کاملSemi-blind Interference Suppression on Coherent Multipath Environments
Blind source separation (BSS) can be used to separate an interfering source from a set of desired signals. By exploiting the statistical independence between the source signals and the interference, we present an alternative to the blind beamformer for suppressing the interference. [1, 2] presented the concept of combining Independent Component Analysis (ICA) with a conventional detection for s...
متن کاملBlind closed-form array response estimation in wireless communication
In this paper, a closed-form array response estimation (CARE) technique for blind source separation in wireless communication is developed. By exploiting the data structure of second-order statistics of the array output in the presence of multipath, we construct a signature matrix in such away that its eigenvectors corresponding to nonezero eigenvalues are just the array response vectors. Thus ...
متن کاملConvolutive Blind Source Separation based on Multiple Decorrelation
| Acoustic signals recorded simultaneously in a reverberant environment can be described as sums of di erently convolved sources. The task of source separation is to identify the multiple channels and possibly to invert those in order to obtain estimates of the underling sources. We tackle the problem by explicitly exploiting the non-stationarity of the acoustic sources. Changing crosscorrelati...
متن کامل