engleski

Signal Denoising using STFT with Bayes prediction and Ephraim-Malah estimation

This paper introduces a new audio cleaning method which is composed of combination of stochastically and orthogonal frequency-based systems. This method can be implemented on signals which have been inherently contaminated with some degree of stationary noise. Beside Short Time Fourier Transform (STFT), this paper focuses on stochastically approach which is needed to ensure the information about minimum mean-square error (MMSE) of the spectral amplitude estimator (SAE). This type of estimation will is performed on a silence- or pause- interval via Bayes prediction method and Ephraim-Malah estimation. This procedure results in with significant reducing of spectral coefficients and therefore the elimination of redundant or noise data. After being performed on an arbitrary mathematical function, the described cleaning method is applied on a PCM (Pulse Code Modulation) wave (22.05 kHz, 8 bit). It is shown that described method is dealing very well with both, noise and discrete disturbance which are the most common problems in the daily work with audio material. The realization of mentioned signal denoising is achieved with MATLAB® developing software

denoising; noise distrubances; audio material; Discrete Fourier Transform (DFT); (STFT); Bayes prediction; Ephraim-Malah estimation; Minimum Mean Square Error (MMSE); spectral amplitude

nije evidentirano