| This page contains example results obtained using the Degenerate Unmixing Estimation Technique (DUET) algorithm as well other BSS results. The DUET algorithm separates an arbitrary number of unknown sources using just two mixtures. The main assumption used is that the time-frequency representations of the sources do not overlap, a condition which we call W-disjoint orthogonality. For anechoic mixtures, a two-dimensional histogram constructed from the ratio of the time-frequency representations of the mixtures is shown to have one peak for each source with peak location corresponding to the relative amplitude and delay mixing parameters. The histogram can be used to partition the time-frequency representation of the mixtures to produce the original sources. |
For information concerning DUET and related techniques, consult the
following papers:
| O. Yilmaz and S. Rickard, Blind Separation of Speech Mixtures via Time-Frequency Masking, IEEE Transactions on Signal Processing, Vol. 52, No. 7, pages 1830-1847, July 2004. [PDF] |
| B. Coleman and S. Rickard, Cardioid Microphones and DUET, IEE Irish Signals and Systems Conference (ISSC2004), pages 264-269, Belfast, Northern Ireland, July 2004. |
| S. Rickard and M. Fallon, The Gini Index of Speech, Proceedings of the 39th Annual Conference on Information Sciences and Systems (CISS2004), Princeton, NJ, March 2004. |
|
S. Rickard, C. Fearon, R. Balan, and J. Rosca, MINUET: Musical Interference Unmixing Estimation Technique, Proceedings of the 39th Annual Conference on Conference on Information Sciences and Systems (CISS2004), Princeton, NJ, March 2004. |
| S. Rickard and O. Yilmaz, On the
W-Disjoint Orthogonality of Speech, IEEE International Conference
on Acoustics, Speech, and Signal Processing (ICASSP2002), Orlando,
Florida, USA, May 13 to 17, 2002. [PDF]
Four source mixing example in Figure 5: Mixtures: x1.wav | x2.wav Demixtures: u1.wav | u2.wav | u3.wav | u4.wav |
| R. Radke and S. Rickard, Audio Interpolation,
International Conference on Virtual, Synthetic and Entertainment Audio (AES22), Espoo, Finland, June 2002. [PDF]
For a demonstration of "Audio Interpolation", consult the following page: Audio Interpolation |
| S. Rickard, R. Balan, and J. Rosca, Real-Time Time-Frequency Based Blind Source Separation, 3rd International Conference on Independent Component Analysis and Blind Source Separation (ICA2001), San Diego, CA, December 9-12, 2001. [PDF] |
| R. Balan, J. Rosca, and S. Rickard, Robustness of Parametric Source Demixing In Echoic Environments, 3rd International Conference on Independent Component Analysis and Blind Source Separation (ICA2001), San Diego, CA, December 9-12, 2001. [PDF] |
| S. Rickard and F. Dietrich, DOA Estimation of Many W-Disjoint Orthogonal Sources from Two Mixtures using DUET, Proceedings of the 10th IEEE Workshop on Statistical Signal and Array Processing (SSAP2000), Pages 311-314, Pocono Manor, PA August 2000. [PDF] |
| A. Jourjine, S. Rickard, and O. Yilmaz, Blind Separation of Disjoint Orthogonal Signals: Demixing N Sources from 2 Mixtures, IEEE Conference on Acoustics, Speech, and Signal Processing (ICASSP2000), Volume 5, Pages 2985-2988, Istanbul, Turkey, June 2000. [PDF] |
| R. Balan, J. Rosca, S. Rickard, and J. O'Ruanaidh, The Influence of Windowing on Time Delay Estimation, Proceedings of the 35th Annual Conference on Information Sciences and Systems (CISS2000), Volume 1, Pages WP1(15-17), Princeton, NJ, March 2000. [PDF] |
For a demonstration of degenerate demixing (5 sources, 2 mixtures), listen to the following:
