Similarity Level Analysis of the Voices of Twins Using the Analysis of Variance and Likelihood Ratio Methods

The Analysis of Variance (ANOVA) and likelihood ratio methods are part of the audio forensic method. These two methods are often used to analyze the level of similarity between the sound recordings of evidence and the voice recordings of the suspect so as to support the legal process. In this study, the ANOVA method and the likelihood ratio were used to analyze the sound similarity of twins. Then an application was built that was built using the calculation of the ANOVA method and the likelihood ratio, where the application can shorten the time in researching sound recordings that have been extracted into formants, and this formant value will later be calculated with the application that has been built. Based on the sample recorded voices of twins that have been analyzed, the results are below 20% for the ANOVA method and below 22% for the likelihood ratio method.