Tài liệu Signal Processing for Telecommunications and Multimedia P1 docx

[...]... (MIMO) systems and space-time coding The last part of the book contains seven chapters that present some emerging system implementations utilizing signal processing to improve system performance and allow for a cost reduction The issues considered range from antenna design and channel equalisation through multi-rate digital signal processing to practical DSP implementation of a wideband direct sequence... the speech and noise sources The algorithm is optimized for finding the speech component in the noisy signal The ability to reduce non-stationary noise sources is investigated 2 FEATURE EXTRACTION FROM SIGNALS The signal of concern is a discrete time noisy speech signal x(n), found from the corresponding correctly band limited and sampled continuous signal It is assumed that the noisy speech signal consists... vectors and the covariance matrices from cepstral domain into the log spectral domain (the indices for state j and mixture k are dropped for simplicity) Equation (1.16) is the standard procedure for linear transformation of a multivariate Gaussian variable Equation (1.17) defines the relationship between the log spectral domain and the linear spectral domain for a multivariate Gaussian variable6 where m and. .. Speech, and Signal Processing, vol ASSP-27(2), pp 113120, April 1979 Deller John R Jr., Hansen John J L., and Proakis John G., Discrete-time processing of speech signals (IEEE Press, 1993, ISBN 0-7803-5386-2) C Jutten and J Heuralt, Blind separation of sources, part i: An adaptive algorithm based on neuromimetic architecture, Signal Processing, vol 24, pp 1-10, June 1991 Y Ephraim, D Malah, and B H... models for enhancing noisy speech, IEEE Transactions on Acoustics, Speech, and Signal Processing, vol 37, no 12, pp 1846-1856, December 1989 H K Kim and R C Rose, Cepstrum-domain model combination based on decomposition of speech and noise for noisy speech recognition, in Proceedings of ICASSP, May 2002, pp 209-212 S J Young and M J F Gales, Cepstral parameter compensation for hmm recognition in noise for. .. Electrical, Computer and Telecommunications Engineering, University of Wollongong, Wollongong, N.S.W 2522, Australia 3 Signal Processing Group, Institute of Physics, University of Oldenburg, 26111 Oldenburg, Germany Abstract We propose a new algorithm for solving the Blind Signal Separation (BSS) problem for convolutive mixing completely in the time domain The closed form expressions used for first and second... the performance of two optimization methods: Gradient, and Newton optimization with speech data Finally, a conclusion is provided in Section 6 The following notations are used in this chapter We use bold upper and lowercase letters to show matrices and vectors, respectively in the time, frequency and domains, e.g., for matrices and for vectors Matrix and vector transpose, complex conjugation, and Hermitian... ESTIMATION 5 The enhancement process, see Fig 1-1, uses information gained from training of speech and noise models to estimate the clean speech signal The information needed are the stored mean vectors in linear spectral domain, and for the speech and the noise respectively, the compensated 8 Chapter 1 model, and the gain difference, g, between training and enhancement environment The stored mean vectors... result from Eq (1.3) The vectors, and are the prototypes for power spectral densities of clean speech and noise respectively Given the compensated model the scaled forward variable, can be found by employing the scaled forward algorithm [10] The scaled forward variable yields the probability vector for being in state j for an observation at time t Given the scaled variable and the mixture weights, it is... vectors, the EM algorithm is applied and the parameters for the HMM are found The model parameter set for an HMM with N states and M mixtures is 1 HMM-Based Speech Enhancement where 5 contains the initial state probabilities, the state transitions probabilities and the parameters for the weighted continuous multidimensional Gaussian functions for state j and mixture k For an observation, the continuous .