The spectral subtractive-type algorithms for enhancing speech in noisy environments

The spectral subtraction is a traditional approach for enhancing the quality of speech degraded by environmental noise. This algorithm is based on the subtraction of the estimated noise spectrum from the noisy speech spectrum and combines it with the phase of the noisy speech. Besides suppressing the noise, this method introduces an unnatural and unpleasant remnant noise. Several variants of this algorithm have been adapted over the years to address its shortcomings, mainly the quality of the remnant noise and its trade-off with speech distortion. This paper presents a comparative performance analysis of different spectral subtractive-type single channel speech enhancement algorithms and experimental results show that the modified forms of spectral subtraction efficiently reduce the remnant noise and improve signal-to-noise ratio.