Speech enhancement using Kalman filtering

in different forms and adopted in various applications. An important parameter of numerous speech enhancement algorithms is the a priori signal-to-noise ratio (SNR). The Wiener filter emphasizes portions of the noisy signal spectrum where SNR is high and attenuates portions of the spectrum where the SNR is low

It is the very essence of our striving for understanding that, on the one hand, it attempts to encompass the great and complex variety of man’s experience, and that on the other, it looks for simplicity and economy in the basic assumptions. The belief that these two objectives can exist side by side is, in view of the primitive state of our scientific knowledge, a matter of faith.

Abstract not found

q Collaboration with industries: v Microsoft v IBM v Intel v Lucent Technologies v Nokia

Speech enhancement is a long standing problem with numerous applications ranging from hearing aids, to coding and automatic recognition of speech signals. In this survey paper we focus on enhancement from a single microphone, and assume that the noise is additive and statistically independent

In this paper, a new approach based on the H1 filtering is presented for speech enhancement. This approach differs from the traditional modified Wiener/Kalman filtering approach in the following two aspects: 1) no aprioriknowledge of the noise statistics is required; instead the noise signals

Abstract—We present a parametric model-based multichannel approach for speech enhancement. By employing an autoregressive model for the speech signal and using a trained codebook of speech linear predictive coefficients, minimum mean square error estimation of the speech signal is performed

Abstract- The speech enhancement problem comprises of various problems characterized by the type of noise source, the nature of interaction between speech and noise, the number of sensor signals (microphone outputs) available for enhancement and the nature of the speech application. Noise reduction

We evaluate some recent developments in recurrent neural network (RNN) based speech enhancement in the light of noise-robust automatic speech recognition (ASR). The proposed framework is based on Long Short-Term Memory (LSTM) RNNs which are discriminatively trained according to an optimal speech