intro algorithm sounds

Blind source separation - algorithm details

Blind source separation refers to the problem of recovering signals from several observed linear (convolutive) mixtures. The strength of the original blind source separation model is that only mutual statistical independence between the source signals is assumed and no a-priori information about,e.g., the characteristics of the source signals, the mixing matrix, or the arrangement of the sensors is needed. In this project, blind source separation is applied to bilateral hearing aids. Microphone signals from both hearing aids are processed simultaneously, leading to a separation of the desired (point) source (e.g., someone you are speaking to) and possible background noise sources (e.g. background music). Thereto, a low-cost, low-delay solution has been derived using frequency-domain-based fast convolution techniques. Similar to current state-of-the-art hearing aids it is assumed that the desired source is located in front of the hearing aid user. This a-priori knowledge about the positions of the microphones allows to obtain a coarse estimate of the positions of the separated sources and makes it possible to select the right output channel containing the desired signal.

References:

Buchner, H., Aichner, R. and Kellermann, W. (2005). A generalization of blind source separation algorithms for convolutive mixtures based on second-order statistics. IEEE Trans. Speech Audio Processing, 13(1):120–134. 

Buchner, H., Aichner, R., Stenglein, J., Teutsch, H. and Kellermann, W. (2005). Simultaneous localization of multiple sound sources using blind adaptive MIMO filtering. In Proc. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2005), Philadelphia, Pennsylvania. 

Aichner, R., Buchner, H. and Kellermann, W., Exploiting Narrowband Efficiency for Broadband Convolutive Blind Source Separation. EURASIP Journal on Applied Signal Processing, Special Issue on Advances in Blind Source Separation, September 2006. 

Aichner, R., Buchner, H., Wehr, S. and Kellermann, W. Robustness of Acoustic Multiple-Source Localization in Adverse Environments. Proc. ITG Fachtagung Sprachkommunikation, Kiel, Germany, April 2006.

Buchner, H., Aichner, R. and Kellermann, W. Blind Source Separation for Convolutive Mixtures: A Unified Treatment. in Audio Signal Processing for Next-Generation Multimedia Communication, Y.Huang, J. Benesty, Eds. Kluwer Academic Publishers, pp. 255-293, 2004.