看板 PhD 關於我們 聯絡資訊
CALL FOR PAPER Special Issue of IEEE Transactions on Audio, Speech and Language Processing **SIGNAL MODELS AND REPRESENTATION OF MUSICAL AND ENVIRONMENTAL SOUNDS** http://www.ewh.ieee.org/soc/sps/tap http://www.ewh.ieee.org/soc/sps/tap/sp_issue/audioCFP.pdf *-- Submission deadline: 15 December, 2008 --* Notification of acceptance: 15 June, 2009 Final manuscript due: 1st July, 2009 Tentative publication date: 1st September, 2009 Guest editors Dr. Bertrand David (Telecom ParisTech, France) bertrand.david@telecom-paristech.fr Dr. Laurent Daudet (UPMC University Paris 06, France) daudet@lam.jussieu.fr Dr. Masataka Goto (National Institute of Advanced Industrial Science and Technology, Japan) m.goto@aist.go.jp Dr. Paris Smaragdis (Adobe Systems, Inc, USA) paris@adobe.com -------------------------------------------------------------------------- The non-stationary nature, the richness of the spectra and the mixing of diverse sources are common characteristics shared by musical and environmental audio scenes. It leads to specific challenges of audio processing tasks such as information retrieval, source separation, analysis-transformation-synthesis and coding. When seeking to extract information from musical or environmental audio signals, the time-varying waveform or spectrum are often further analysed and decomposed into sound elements. Two aims of this decomposition can be identified, which are sometimes antagonist: to be together adapted to the particular properties of the signal and to the targeted application. This special issue is focused on how the choices of a low level representation (typically a time-frequency distribution with or without probabilistic framework, with or without perceptual considerations), a source model or a decomposition technique may influence the overall performance. Specific topics of interest include but are not limited to: * factorizations of time-frequency distribution * sparse representations * Bayesian frameworks * parametric modeling * subspace-based methods for audio signals * representations based on instrument or/and environmental sources signal models * sinusoidal modeling of non-stationary spectra (sinusoids, noise, transients) Typical applications considered are (non exclusively): * source separation/recognition * mid or high level features extraction (metrics, onsets, pitches, …) * sound effects * audio coding * information retrieval * audio scene structuring, analysis or segmentation * ... -- ※ 發信站: 批踢踢實業坊(ptt.cc) ◆ From: 218.175.217.142