Refinement of input speech by suppressing the unwanted amplitudes for blue hearing system

In this paper we have developed the mechanism to remove unwanted low or high amplitudes from the given input speech. The input speech is divided into many small segments of very short interval. These segments are checked against the given amplitude. The unwanted amplitude segments are filtered out while the desired amplitude segments are selected. All selected segments are combined to produce the output. The comparison of accepted and rejected segments for various cases is very much significant to measure the percentage of input speech lying between two amplitude values. The hearing dimensions of the deaf patients vary so this mechanism can be imposed in the speech processor of Blue Hearing System, which is an artificial hearing system for severely deaf or inhospitably hard of hearing people. The proposed algorithm and all experimental works have been implemented in MATLAB.