Multi-Channel Estimation of the Power Spectral Density of Noise for Mixtures of Non-Stationary Signals
نویسندگان
چکیده
The proposed paper deals with the estimation of the power spectral density (PSD) of noise for mixtures of non-stationary wideband signals exploiting sparseness in the time-frequency domain. The proposed method is applied to realize a beamformer-plus-postfilter structure for noise-robust speech recognition. Experiments with a small-scale 4-sensor microphones array show that interfering speech and music can be suppressed by 13 dB for an average signal-to-noise ratio of 0 dB without impairing the quality of the desired speech signal and without introducing ’musical noise’. This noise suppression is reflected by the improvement of the word accuracy of a connected-digit speech recognizer.
منابع مشابه
IMPLEMENTATION OF EXTENDED KALMAN FILTER TO REDUCE NON CYCLO-STATIONARY NOISE IN AERIAL GAMMA RAY SURVEY
Gamma-ray detection has an important role in the enhancement the nuclear safety and provides a proper environment for applications of nuclear radiation. To reduce the risk of exposure, aerial gamma survey is commonly used as an advantage of the distance between the detection system and the radiation sources. One of the most important issues in aerial gamma survey is the detection noise. Various...
متن کاملEstimation of harmonic interference parameters of surface-NMR signal using an adaptive method and residual signal power
Surface nuclear magnetic resonance (surface-NMR) method is a well-known tool for determining the water-bearing layers and subsurface resistivity structure. Harmonic interference is an inevitable interference in surface-NMR measurements. Accurate estimation of harmonic interference parameters (i.e., fundamental frequency, phase and amplitude) leads to better retrieval of power-line harmonics and...
متن کاملSpectral Estimation of Stationary Time Series: Recent Developments
Spectral analysis considers the problem of determining (the art of recovering) the spectral content (i.e., the distribution of power over frequency) of a stationary time series from a finite set of measurements, by means of either nonparametric or parametric techniques. This paper introduces the spectral analysis problem, motivates the definition of power spectral density functions, and reviews...
متن کاملGeneralized Eigenvector Blind Speech Separation under Coherent Noise in a Gsc Configuration
This paper deals with a new technique for multi-channel separation of speech signals from convolutive mixtures under coherent noise. We demonstrate how the scaled transfer functions from the sources to the microphones can be estimated even in the presence of stationary coherent noise. The key to this are generalized eigenvalue decompositions of the power spectral density (PSD) matrices of the n...
متن کاملDual Channel Reduction of Rapidly Varying Harmonic and Random Noise Using a Spot Microphone
This contribution presents a dual channel speech enhancement system that is operating in the frequency domain. State-of-the-art estimators usually rely on the assumption that the noise signal is stationary or only slightly time-varying. In this contribution we are explicitly considering rapidly time-varying harmonic noise (e.g., engine noise). Therefore we take a spot microphone into account en...
متن کامل