Simultaneous Detection and Estimation Approach in Speech Enhancement using Weighted Cosh Estimator
نویسندگان
چکیده
Traditional speech enhancement algorithms consider speech detection and estimation process independently. The simultaneous detection and estimation approach (SDA) combines both detection and estimation processes to reduce the effects due to missed detection. SDA is based on an estimator which minimizes bayesian risk with the aid of minimum mean square error (MMSE) estimator. The MMSE estimator need not necessarily consider the masking effects and the appropriate weight that has to be given for the errors at spectral valleys to remove the residual noise. Hence the conventional MMSE is replaced with the weighted COSH (WCOSH) estimator in our approach, which is based on the perceptually motivated cost functions. We proposed SDA-WCOSH estimator by incorporating WCOSH into the SDA which minimizes the cost function that takes into account of both detection and estimation errors. The proposed SDA-WCOSH performed better than MMSE, STSA, SDA algorithms and it was successful in removing the residual noise in a better manner. Our analysis also showed that the SDA-WCOSH algorithm also resulted in the lesser speech distortion during enhancement when compared to SDA and STSA.
منابع مشابه
Weighted Log-spectral Amplitude Estimation with Generalized Gamma Distribution under Speech Presence Probability
In this paper, we propose a speech enhancement approach. The approach is based on deriving weighted log-spectral amplitude estimator that exploits the generalized Gamma distributed speech priors under speech presence probability. The log-spectral amplitude estimator is weighted by psychoacoustically motivated speech distortion measure to take advantage of the perceptual interpretation. The expe...
متن کاملSimultaneous robust estimation of multi-response surfaces in the presence of outliers
A robust approach should be considered when estimating regression coefficients in multi-response problems. Many models are derived from the least squares method. Because the presence of outlier data is unavoidable in most real cases and because the least squares method is sensitive to these types of points, robust regression approaches appear to be a more reliable and suitable method for addres...
متن کاملSpeech enhancement based on hidden Markov model using sparse code shrinkage
This paper presents a new hidden Markov model-based (HMM-based) speech enhancement framework based on the independent component analysis (ICA). We propose analytical procedures for training clean speech and noise models by the Baum re-estimation algorithm and present a Maximum a posterior (MAP) estimator based on Laplace-Gaussian (for clean speech and noise respectively) combination in the HMM ...
متن کاملSpeech Enhancement using Laplacian Mixture Model under Signal Presence Uncertainty
In this paper an estimator for speech enhancement based on Laplacian Mixture Model has been proposed. The proposed method, estimates the complex DFT coefficients of clean speech from noisy speech using the MMSE estimator, when the clean speech DFT coefficients are supposed mixture of Laplacians and the DFT coefficients of noise are assumed zero-mean Gaussian distribution. Furthermore, the MMS...
متن کاملSpeech Enhancement Using Gaussian Mixture Models, Explicit Bayesian Estimation and Wiener Filtering
Gaussian Mixture Models (GMMs) of power spectral densities of speech and noise are used with explicit Bayesian estimations in Wiener filtering of noisy speech. No assumption is made on the nature or stationarity of the noise. No voice activity detection (VAD) or any other means is employed to estimate the input SNR. The GMM mean vectors are used to form sets of over-determined system of equatio...
متن کامل