This page shows demo of our proposed algorithms submitted to ICASSP 2022 and other blind source separation algorithms.
Paper title: HARVESTING PARTIALLY-DISJOINT TIME-FREQUENCY INFORMATION FOR IMPROVING DEGENERATE UNMIXING ESTIMATION TECHNIQUE
Authors: Yudong He1,2, He Wang2, Qifeng Chen1 , Richard H.Y. So1,2
- 1 The Hong Kong University of Science and Technology, Hong Kong, China
- 2 HKUST-Shenzhen Research Institute, Shenzhen, China
Matlab code is availabel!
We have open-sourced our Matlab codes for all blind source algorithms that we used in our paper. Please find them here
Two mixtures of two sources example
Original source signals
Live recorded mixtures (reverb=130ms)
Separated audio signals by different blind source separation algorithms
Live recorded mixtures (reverb=250ms)
Separated audio signals by different blind source separation algorithms
Two mixtures of three sources example
Original source signals
Live recorded mixtures (reverb=130ms)
Separated audio signals by different blind source separation algorithms
Live recorded mixtures (reverb=250ms)
Separated audio signals by different blind source separation algorithms
Two mixtures of four sources example
Original source signals
Live recorded mixtures (reverb=130ms)
Separated audio signals by different blind source separation algorithms
Live recorded mixtures (reverb=250ms)
Separated audio signals by different blind source separation algorithms
Computational time
| Algorithms |
2 sources |
3 sources |
4 sources |
| DUET-MVDR (proposed) |
0.74 s |
0.84 s |
0.96 s |
| DUET-ISR (proposed) |
0.73 s |
1.00 s |
1.01 s |
| DUET [1] |
0.10 s |
0.14 s |
0.18 s |
| IVA [2] |
3.40 s |
\ |
\ |
| ILRMA [3] |
8.39 s |
\ |
\ |
| MULTI_NMF [4] |
\ |
32.85 s |
43.54 s |
| FULLRANK [5] |
\ |
575.32 s |
585.39 s |
Reference
- [1] Scott Rickard, The DUET Blind Source Separation Algorithm, pp. 217–241, Springer Netherlands, Dordrecht, 2007.
- [2] Taesu Kim, Torbjørn Eltoft, and Te-Won Lee, “Independent vector analysis: An extension of ica to multivariate components,” in Independent Component Analysis and Blind Signal Separation, Justinian Rosca, Deniz Erdogmus, José C. Prı́ncipe, and Simon Haykin, Eds., Berlin, Heidelberg, 2006, pp. 165–172, Springer Berlin Heidelberg.
- [3] Daichi Kitamura, Nobutaka Ono, Hiroshi Sawada, Hirokazu Kameoka, and Hiroshi Saruwatari, “Determined blind source separation unifying independent vector analysis and nonnegative matrix factorization,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 24, no. 9, pp. 1626–1641, 2016.
- [4] A. Ozerov and C. Fevotte, "Multichannel nonnegative matrix factorization in convolutive mixtures for audio source separation," IEEE Trans. on Audio, Speech and Lang. Proc. special issue on Signal Models and Representations of Musical and Environmental Sounds, vol. 18, no. 3, pp. 550-563, March 2010.
- [5] N. Q. K. Duong, E. Vincent and R. Gribonval, "Under-determined reverberant audio source separation using a full-rank spatial covariance model", IEEE Transactions on Audio, Speech and Language Processing, 2010