Speech enhancement using minimum mean square error filter and spectral subtraction filter

Speech, is very important for us to communicate. In speech communication, the speech signal is always accompanied by some noise. Hence speech enhancement technique is used for the removal of unwanted background noise. The objective of speech enhancement is to improve the perceptual aspects of speech such as overall quality, intelligibility, or degree of listener´s fatigue. The goal of this work is to enhance the quality of speech signals using combination of filters to improve the intelligibility of the signal. In this work, a combination of spectral subtraction filter and minimum mean square error filter is used for enhancing the noisy speech signal. The performance of the proposed filter combination is evaluated using Noizeus database. The signal-to-noise ratio and log-likelihood ratio metrics of the proposed filter shows the increase in performance.