کنفرانس مهندسی برق ایران

صفحه اصلی / سی امین کنفرانس بین المللی مهندسی برق

Instantaneous Blind Audio Source Separation Using Characteristic Function of Heavy-Tailed Distributions

نویسندگان :

Kamran Rajabi¹ Mohammadreza Hassannejad Bibalan² Neda Faraji³

1- دانشگاه بین‌المللی امام خمینی (ره) 2- دانشگاه بین‌المللی امام خمینی (ره) 3- دانشگاه بین‌المللی امام خمینی (ره)

کلمات کلیدی :

Blind audio source separation،symmetric alpha-stable distributions،characteristic function،linear instantaneous model.

چکیده :

Blind audio source separation (BASS) methods separate combined audio signals up to unknown scaling and ordering factors. In this study by assigning symmetric alpha-stable (S\textalpha S) distributions to sources interesting approach is proposed. Considering the fact that S\textalpha S distributions can better model audio signal due to their heavier tails, our method outperformed other well-renowned algorithms. The main drawback of S\textalpha S distributions is the lack of closed-form expression for its density function, apart from some exceptional cases. But promising fact is that characteristic function of it has closed-formed relation. To tackle BASS problem, discrete formula of characteristic function, which was defined based on spectral measure, is used. Our signals are combined linearly and instantaneously in each sensor, number of sources and sensors are equal, which is so-called determined case. Notwithstanding, our method can be implemented in determined case, it is proven method has its own efficiency even in mild under-determined case. Devising solution for BASS problem, sources are presumed to be independent and identically distributed (i.i.d.), which is not far from actual situations.