An automatic grading method for singing evaluation

Abstract

Automatic evaluation of singing quality is essential to singing practitioners. In this paper, we present a methodology that generates grades and comments for singing quality. In this method, we first perform note segmentation and pitch estimation. Then a method based on experience of human perception is used to evaluate the accuracies of intonation and rhythm. A subjective evaluation is carried out in which Spearmans correlation is used to evaluate the consistency between the proposed method and experts opinion. The result demonstrates high reliability of the proposed annotation system and grading method. © 2012 Springer-Verlag Berlin Heidelberg.