The Comparison of Audio Analysis Using Audio Forensic Technique and Mel Frequency Cepstral Coefficient Method (MFCC) as the Requirement of Digital Evidence

Abstrak

Audio forensics is the application of science and scientific methods in handling evidence in the form of audio to support the disclosure of various criminal cases and to reveal information needed in the trial process. In this regard, a sound recording that has been made is then cut into words. Then, pieces of it were analyzed by using audio forensic techniques through parameters of pitch, formant and spectogram using Forensic Audio Technique Analysis on artificial voice recordings and live voice recordings. The analysis was also carried out using the extraction of the Mel Frequency Cepstral Coefficient (MFCC) feature, the Dynamic Time Warping (DTW) Method, and applying the K-Nearest Neighbor (KNN) algorithm. The testing results by using audio forensic techniques obtained an accuracy of 76.3%, meanwhile the accuracy of testing results by using a system that has been built (self-made system) is 66.7%.