• Monish Khandelwal

Speech enhancement with an objective to suppress noise can be a complex process and cause distortions or loss of information while processing the degraded speech signal. Enhancement with the help of Power spectrums (such as SNR) is used for processing speech signals in Spectral domain. The basic and the oldest algorithms proposed in this domain were the Spectral Subtraction and Mean Minimum Square Error. The present paper addresses detailed process of Spectral Subtraction and MMSE algorithms with a mathematical and practical approach. Spectral Subtraction is simple for implementation but suffers from distortions which are sometimes generated due to uncontrolled processing of the signals. However, this distortion can be reduced by making different approach to eliminate the noise. The MMSE algorithms are basically statistical estimators in which different assumptions and estimation are made required to suppress the noise present in the noisy speech signals. The consistency, drawbacks and results are too analyzed in the paper.

Speech Enhancement, Spectral Processing, Spectral Subtraction, Mean Minimum Square Error, Maximum a posteriori, Zero Cross-terms.

"SUPPRESSION OF ADDITIVE BACKGROUND NOISE FROM DEGRADED SPEECH SIGNALS BY SPECTRAL PROCESSING", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 6, page no.872-876, June-2019, Available :http://www.jetir.org/papers/JETIR1906130.pdf

"SUPPRESSION OF ADDITIVE BACKGROUND NOISE FROM DEGRADED SPEECH SIGNALS BY SPECTRAL PROCESSING", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 6, page no. pp872-876, June-2019, Available at : http://www.jetir.org/papers/JETIR1906130.pdf