Academic
Publications
Blind separation of convolved mixtures in the frequency domain

Blind separation of convolved mixtures in the frequency domain,10.1016/S0925-2312(98)00047-2,Neurocomputing,Paris Smaragdis

Blind separation of convolved mixtures in the frequency domain   (Citations: 344)
BibTex | RIS | RefWorks Download
In this paper we employ information theoretic algo- rithms, previously used for separating instantaneous mixtures of sources, for separating convolved mix- tures in the frequency domain. It is observed that convolved mixing in the time domain corresponds to instantaneous mixing in the frequency domain. Such mixing can be inverted using simpler and more robust algorithms than the ones recently developed. Advantages of this approach are improved efficiency and better convergence features.
Journal: Neurocomputing - IJON , vol. 22, no. 1-3, pp. 21-34, 1998
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
    • ...problem on time domain [8] and on frequency domain [9]...
    • ...frequency domain, can deal with convolved mixture data [8][9]...

    Viet-Hung Danget al. Clustering based multi-object positioning system

    • ...To solve this problem, sound source separation has been extensively studied (e.g., [2, 3, 4, 5])...
    • ...Some of them used independent component analysis (ICA) [2, 3] and others employed binary masks in the spectrogram [4, 5]. The ICA-based methods, however, need high computational costs to calculate higher-order statistics...

    Yu Nasuet al. Cross-Channel Spectral Subtraction for meeting speech recognition

    • ...Although the frequency-domain approach for blind separation of convolutive speech mixtures has been studied since 1990s [5, 20], and although the properties of the timedomain real-valued speech have been studied in details from the view of BSS [5, 17, 18, 20], the characteristics of the frequency-domain complex-valued sources are not thoroughly investigated to validate whether the assumption required by different complex-valued BSS ...
    • ...Although the frequency-domain approach for blind separation of convolutive speech mixtures has been studied since 1990s [5, 20], and although the properties of the timedomain real-valued speech have been studied in details from the view of BSS [5, 17, 18, 20], the characteristics of the frequency-domain complex-valued sources are not thoroughly investigated to validate whether the assumption required by different complex-valued BSS ...
    • ...Another important issue to the frequency-domain blind separation of convolutive speech is the duration of the sliding window [5, 17, 18, 20]...

    Fengyu Conget al. Second order impropriety based complex-valued algorithm for frequency-...

    • ...In order to mitigate such difficulties, frequency-domain BSS methods were proposed, where convolutive mixtures can be treated as instantaneous mixtures in each frequency bin [3], [4]...

    Paulo Bulkool Batalheiroet al. Subband blind source separation considering acoustic reverberation cha...

    • ...Many approaches have been developed to solve this problem both in time [Torkkola, 1996; Lee et al., 1997] and frequency domains [Ciaramella et al., 2006; Smaragdis, 1998]...

    A. Ciaramellaet al. Automatic detection of long-period events at Campi Flegrei Caldera (It...

Sort by: