Blind Speech Separation

ISBN-13: 9781402064784 / Angielski / Twarda / 2007 / 432 str.

Te-Won Lee; Hiroshi Sawada; S. Makino

Opis

We are surrounded by sounds. Such a noisy environment makes it di?cult to obtain desired speech and it is di?cult to converse comfortably there. This makes it important to be able to separate and extract a target speech signal from noisy observations for both man machine and human human communication. Blindsourceseparation(BSS)isanapproachforestimatingsourcesignals using only information about their mixtures observed in each input channel. The estimation is performed without possessing information on each source, such as its frequency characteristics and location, or on how the sources are mixed. The use of BSS in the development of comfortable acoustic com- nication channels between humans and machines is widely accepted. Some books have been published on BSS, independent component ana- sis (ICA), and related subjects. There, ICA-based BSS has been well studied in the statistics and information theory ?elds, for applications to a variety of disciplines including wireless communication and biomedicine. However, as speech and audio signal mixtures in a real reverberant environment are generally convolutive mixtures, they involve a structurally much more ch- lenging task than instantaneous mixtures, which are prevalent in many other applications."

Spis treści

Part I: Multiple Microphone Blind Speech Separation with ICA 1. Convolutive Blind Source Separation for Speech Signals; S.C.Douglas, M.Gupta. 2. Frequency-Domain Blind Source Separation; H.Sawada, S.Araki, S.Makino. 3. Blind Source Separation using Space-Time Independent Component Analysis; M.Davies, et al. 4. TRINICON-based Blind System Identification with Application to Multiple-Source Localization and Separation; H.Buchner, R.Aichner, W.Kellermann. 5. SIMO-Model-Based Blind Source Separation-principle and its applications; H.Saruwatari, T.Takatani, K.Shikano. 6. Independent Vector Analysis for Convolutive Blind Speech Separation; I.Lee, T.Kim, T-W.Lee. 7. Relative Newton and Smoothing Multiplier Optimization Methods for Blind Source Separation; M.Zibulevsky. Part II: Underdeterminded Blind Speech Separation with Sparseness 8. The DUET Blind Source Separation Algorithm; S.Rickard. 9. K-means Based Underdetermined Blind Speech Separation; S.Araki, H.Sawada, S.Makino. 10. Underdetermined Blind Source Separation of Convolutive Mixtures by Hierarchical Clustering and L1-Norm Minimization; S.Winter, et al. 11. Bayesian Audio Source Separation; C.Févotte. Part III: Single Microphone Blind Speech Separation 12. Monaural Source Separation; G.J.Jang, T-W.Lee. 13. Probabilistic Decompositions of Spectra for Sound Separation; P.Smaragdis. 14. Sparsification for Monaural Source Separation; H.Asari, et al. 15. Monaural Speech Separation by Support Vector Machines; S.Hochreiter, M.C.Mozer. Index.

O autorze

Dr. Shoji Makino is an IEEE Fellow, Associate Editor of the IEEE Transactions on Speech & Audio Processing, and Executive Manager NTT Communication Science Laboratories. Dr. Makino was also co-editor on the succesful 2005 Springer book: Benesty - Speech Enhancement.

Blurb

This is the first book to provide a cutting edge reference to the fascinating topic of blind source separation (BSS) for convolved speech mixtures. Through contributions by the foremost experts on the subject, the book provides an up-to-date account of research findings, explains the underlying theory, and discusses potential applications. The individual chapters are designed to be tutorial in nature with specific emphasis on an in-depth treatment of state of the art techniques.

Blind Speech Separation is divided into three parts:

Part 1 presents overdetermined or critically determined BSS. Here the main technology is independent component analysis (ICA). ICA is a statistical method for extracting mutually independent sources from their mixtures. This approach utilizes spatial diversity to discriminate between desired and undesired components, i.e., it reduces the undesired components by forming a spatial null towards them. It is, in fact, a blind adaptive beamformer realized by unsupervised adaptive filtering.

Part 2 addresses underdetermined BSS, where there are fewer microphones than source signals. Here, the sparseness of speech sources is very useful; we can utilize time-frequency diversity, where sources are active in different regions of the time-frequency plane.

Part 3 presents monaural BSS where there is only one microphone. Here, we can separate a mixture by using the harmonicity and temporal structure of the sources. We can build a probabilistic framework by assuming a source model, and separate a mixture by maximizing the a posteriori probability of the sources.