Joint spatial filter and time-varying MCLP for dereverberation and interference suppression of a dynamic/static speech source
Dereverberation of a moving speech source in the presence of other directional interferers, is a harder problem than that of stationary source and interference cancellation. We explore joint multi channel linear prediction (MCLP) and relative transfer function (RTF) formulation in a stochastic framework and maximum likelihood estimation. We found that the combination of spatial filtering with distortion-less response constraint, and time-varying complex Gaussian model for the desired source signal at a reference microphone does provide better signal estimation. For a stationary source, we consider batch estimation, and obtain an iterative solution. Extending to a moving source, we formulate a linear time-varying dynamic system model for the MCLP coefficients and RTF based online adaptive spatial filter. For the case of tracking a desired source in the presence of interfering sources, the same formulation is used by specifying the RTF. Simulated experimental results show that the proposed scheme provides better spatial selectivity and dereverberation than the traditional methods, for both stationary and dynamic sources even in the presence of interfering sources.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Directional MCLP Analysis and Reconstruction for Spatial Speech Communication
Spatial speech communication, i.e., the reconstruction of spoken signal along with the relative speaker position in the enclosure (reverberation information) is considered in this paper. Directional, diffuse components a…
PositionAn Effective Dereverberation Algorithm by Fusing MVDR and MCLP
In the scenario with reverberation, the experience of human-machine interaction will become worse. In order to solve this problem, many methods for the dereverberation have emerged. At present, how to update the paramete…
BlockingMCLPD:Multi-view Contrastive Learning for EEG-based PD Detection Across Datasets
Electroencephalography has been validated as an effective technique for detecting Parkinson's disease,particularly in its early stages.However,the high cost of EEG data annotation often results in limited dataset size an…
Self-Supervised LearningContrastive LearningElectric vehicle demand estimation and charging station allocation using urban informatics
This paper performs a novel data-driven approach to optimize electric vehicle (EV) public charging. We translate the study area into a directed graph by partitioning it into discrete grids. A modified geographical Page…
Exploiting spatial information with the informed complex-valued spatial autoencoder for target speaker extraction
In conventional multichannel audio signal enhancement, spatial and spectral filtering are often performed sequentially. In contrast, it has been shown that for neural spatial filtering a joint approach of spectro-spatial…
PositionTarget Speaker Extraction