IEEE Transactions on Speech and Audio Processing · 2005 · 336 citations · 25 references
Health ProfessionsAudiologyAcousticsComputational AcousticsPsychoacousticsDigital Signal ProcessingNoiseSpeech AcousticsSpeech IntelligibilitySpeech Intelligibility EnhancementSpeech RecognitionDiscrete Fourier TransformDiscrete-time Fourier TransformProbability Density FunctionsSpeech EnhancementSpeech Signal AnalysisSpeech Signal EnhancementSpeech Signal ProcessingWiener Filtering
The paper introduces a class of minimum mean‑square error (MMSE) estimators for enhancing short‑time spectral coefficients of a noisy speech signal. The authors derive analytical MMSE estimators for DFT coefficients that employ super‑Gaussian priors—complex Laplace or bilateral Gamma for clean speech and complex Gaussian or Laplacian for noise—rather than the conventional Gaussian assumption. These super‑Gaussian‑based estimators achieve a higher signal‑to‑noise ratio than Gaussian‑assumed algorithms such as the Wiener filter or the Ephraim‑Malah MMSE spectral amplitude estimator.
This paper presents a class of minimum mean-square error (MMSE) estimators for enhancing short-time spectral coefficients of a noisy speech signal. In contrast to most of the presently used methods, we do not assume that the spectral coefficients of the noise or of the clean speech signal obey a (complex) Gaussian probability density. We derive analytical solutions to the problem of estimating discrete Fourier transform (DFT) coefficients in the MMSE sense when the prior probability density function of the clean speech DFT coefficients can be modeled by a complex Laplace or by a complex bilateral Gamma density. The probability density function of the noise DFT coefficients may be modeled either by a complex Gaussian or by a complex Laplacian density. Compared to algorithms based on the Gaussian assumption, such as the Wiener filter or the Ephraim and Malah (1984) MMSE short-time spectral amplitude estimator, the estimators based on these supergaussian densities deliver an improved signal-to-noise ratio.
25
Table of Integrals, Series, and Products.
K. S. Kölbig, I. S. Gradshteyn, I.M. RYZHIK et al. · Mathematics of Computation · 1995 · 9.6K citations
Continuous Univariate Distributions.
Yasuhiro Omori, Norman L. Johnson, S. Kotz et al. · Journal of the American Statistical Association · 1995 · 9.3K citations
Information theory and statistics
Journal of the Franklin Institute · 1959 · 7.2K citations
Ralph G. Hudson · Medical Entomology and Zoology · 1917 · 7K citations
Natural Sciences, Computer Science, Differential Equations +7