Noise Removal in Speech Signal using Modified Berouti Spectral Subtraction for Emotion Recognition

In this paper, Berouti Spectral subtraction is used for reducing noise from noisy speech signals. It calculates the spectrum of the noisy speech using the combination of Fast Fourier Transform (FFT) and spectral flux and then the noise spectrum is subtracted from the noisy speech spectrum. The performance of this paper was measured by calculating the Signal to Noise Ratio (SNR). This paper proposes a new parameter for speech enhancement. The idea is to apply the spectral subtraction with spectral flux for estimating the noise more precisely. The new parameter is used for spectral subtraction in unvoiced speech frames and the existing power factor in spectral subtraction method is improved.