IEEE Transactions on Audio Speech and Language Processing · 2007 · 235 citations · 25 references
EngineeringSpeech-language PathologyAcousticsAntennasComputational AcousticsArray ProcessingBeamformingDigital Signal ProcessingAcoustic BeamformingBeamforming TechniquesConvergenceEigenvalue ComputationMinimum Variance Distortionless Response BeamformingSignal-to-noise RatioSignal-to-noise Ratio Estimation
Maximizing output SNR in sensor arrays leads to a generalized eigenvalue problem, a technique common in narrowband antenna arrays but rarely used for broadband microphone arrays because a narrowband SNR criterion introduces uncontrolled speech distortion. This study demonstrates that a single‑channel post‑filter can control desired‑signal distortion, achieving performance comparable to a generalized minimum‑variance distortionless response beamformer. The proposed blind beamformer employs generalized eigenvalue decomposition without requiring array geometry, explicit transfer‑function estimation, or direction‑of‑arrival knowledge, and introduces a gradient‑ascent adaptation algorithm. Experimental results show the method performs well under both directional and diffuse noise, and the gradient‑ascent algorithm converges faster than existing alternatives.
Maximizing the output signal-to-noise ratio (SNR) of a sensor array in the presence of spatially colored noise leads to a generalized eigenvalue problem. While this approach has extensively been employed in narrowband (antenna) array beamforming, it is typically not used for broadband (microphone) array beamforming due to the uncontrolled amount of speech distortion introduced by a narrowband SNR criterion. In this paper, we show how the distortion of the desired signal can be controlled by a single-channel post-filter, resulting in a performance comparable to the generalized minimum variance distortionless response beamformer, where arbitrary transfer functions relate the source and the microphones. Results are given both for directional and diffuse noise. A novel gradient ascent adaptation algorithm is presented, and its good convergence properties are experimentally revealed by comparison with alternatives from the literature. A key feature of the proposed beamformer is that it operates blindly, i.e., it neither requires knowledge about the array geometry nor an explicit estimation of the transfer functions from source to sensors or the direction-of-arrival.
25
High-resolution frequency-wavenumber spectrum analysis
J. Capon · Proceedings of the IEEE · 1969 · 6.2K citations