Speech enhancement is concerned with the processing of corrupted or noisy speech signal in order to improve the quality or intelligibility of the signal. There are so many applications of speech still to be far from reality just because of lack of efficient and reliable noise removal mechanism and preserving or improving the intelligibility for the speech signals. The aim of the speech enhancem...

in this paper, we propose a novel algorithm to enhance the noisy speech in the framework of dual-channel speech enhancement. the new method is a hybrid optimization algorithm, which employs the  combination of  the  conventional θ-pso and the shuffled sub-swarms particle optimization (sspso) technique. it is known that the θ-pso algorithm has better optimization performance than standard pso al...

Recognition rate of noisy short utterance is lower, the two main factors are the inadequate training data and utterance polluted by noisy seriously. In this paper, we proposed corresponding algorithms. First, noise and speech are regarded as parallel information, we use FastICA algorithm to separate pure speech and noise. And then, we use differences detecting and eliminating algorithm (DDAEA) ...

A noise-estimation algorithm is proposed for highly non-stationary noise environments. The noise estimate is updated by averaging the noisy speech power spectrum using time and frequency dependent smoothing factors, which are adjusted based on signal-presence probability in individual frequency bins. Signal presence is determined by computing the ratio of the noisy speech power spectrum to its ...

Compressed sensing (CS) is a sampled approach on signal sparsity-base, and it can effectively extract the information which is contained in the signal. This paper presents a noisy speech enhancement new method based on CS process. Algorithm uses a voice sparsity in the discrete fast Fourier transform (Fast Fourier transform, FFT), and complex domain observation matrix is designed, and the noisy...

The report deals with speech enhancement in a noisy environment. A new, direct shaping parametric formulation of generalised adaptive spectral subtraction algorithm for non-stationary interference is considered. The idea of the authors is not to select but to form desirable envelope of spectral estimator function in SNR*Gain-function area. Some principles of parameters choice are proposed, base...

This paper proposes the application of tree-structured clustering to various noise samples or noisy speech in the framework of piecewise-linear transformation (PLT)-based noise adaptation. According to the clustering results, a noisy speech HMM is made for each node of the tree structure. Based on the likelihood maximization criterion, the HMM that best matches the input speech is selected by t...

A speech denoising technique based on subband noise estimation and a perceptual modification of Wiener filtering is proposed. The noisy speech is first decomposed into critical band signals by an auditory filterbank and the denoising is carried out on the subband signals. The time varying subband noise variance required for denoising is estimated by tracking the minimum variance of the subband ...

We investigate methods of improving the intelligibility of synthetic speech under noisy or low-fidelity acoustic conditions. Techniques explored improve speech in a natural manner, such that training won’t be required for the user to understand the enhanced speech. While the improvements are natural in this respect, the changes aren’t limited to creating only speech that is achievable by a huma...

The quality of text-to-speech (TTS) voices built from noisy speech is compromised. Enhancing the speech data before training has been shown to improve quality but voices built with clean speech are still preferred. In this paper we investigate two different approaches for speech enhancement to train TTS systems. In both approaches we train a recursive neural network (RNN) to map acoustic featur...