Separating 3 Speech Signals using 2
Microphones for Convolutive Mixtures (DEMO PAGE)
Alireza
Masnadi-Shirazi, Wenyi
Zhang and Bhaskar D. Rao
Real room mixtures:
1.
Recorded in a conference room setting
2.
The sources were 3 loudspeakers positioned on
a table, 1-2 m away from microphone pair
3.
Distance between the 2 microphones was 4cm
Simulated room mixtures:
1.
Used Image method to simulate room mixing
2.
1.5 m away from microphone pair
3.
Distance between the 2 microphones was 10 cm
Audio Files (Male, Digits, Real room): (if files do not play when clicked, right
click and save to your computer)
True source 1 True
source 2 True source 3
Glimpsing-IVA:
Separated source 1 Separated source 2 Separated source 3
Sawada
et al. T-F masking[1]: Separated source 1 Separated source 2 Separated source 3
Audio Files (Male, Digits, Simulated room):
True source 1 True
source 2 True source 3
Glimpsing-IVA:
Separated source 1 Separated
source 2 Separated source 3
Sawada
et al. T-F masking[1]: Separated
source 1 Separated source 2 Separated source 3
Audio Files (Female, Real room):
True source 1 True source 2 True source 3
Glimpsing-IVA:
Separated source 1 Separated
source 2 Separated source 3
[1] H. Sawada, S. Araki, and S.
Makino, A two stage frequency-domain blind source separation method for
underdetermined
convolutive
mixtures, in 2007 IEEE Workshop on Applications of Signal Processing to
Audio and Acoustics, 2007