AI-Based Vocal Judging Application

Abstract

Some contesters in musical contests and singing competitions are lacking a proper coach and there may be misjudgments by the judges of the competitions. To overcome that it is possible to apply AI and Deep Learning and get a computer to give an average reasonable score for the performance by analyzing the frequencies. In this research, A possible solution for this occasion is described. The objective of this research is to help the judges or the individual users to give a fair and more accurate grading for the songs. The introduced solution uses FFT Algorithm as the main method to analyze the frequencies. And this will get a final average grade by analyzing, Pitch, Musical Notations and Tempo of the performance. The accuracy of the analyzing system is about 97% at the current time. This may be varied by the selection of separation, Frequency analyzing, and Midi note detection algorithms. These detections are evaluated by comparing it to FL Studio which is a software that is available for professional music development.