The Utilization of Gaussian Filter Method on Voice Record Frequency Noise

Abstract

Voice is one of the ubiquitous media and is commonly used as communication by humans with the hope that it must reach the destination clearly and can be understood by those who receive the voice, but cannot support the delivery of voice information, because in sending information on the sound provided by the noise speaking the arrival of sound. One solves the problem of sound signals disturbed by noise, which can be facilitated by digital filters. The Gaussian method is the best method for eliminating noise in the image developed so that the gaussian method needs to be done to improve the noise in the sound to issue a clean sound signal from noise and increase the noise that has been eliminated. This study uses the Gaussian filter method to filter noise in the sound record. The implementation is divided into two (2) steps to filter on the frequency consisting of images and the second to do with voice recording, in carrying out the filtering on the frequency of the sound that is considered noise. In filtering the Gaussian method to eliminate the noise of the sound record, the current process is combined with obtaining a dynamic range, and peak factor with parameters attack time = 0.01, release time = 0.01, hold time = 0.05, lower limit = 0.0005, and threshold of silence = 0.25. Apart from the results in this paper, the dynamic range and crest factors in the filtering parameters require a very long process, and an optimal method of repair is needed.