Music Tutor: Application of Chord Recognition in Music Teaching

Abstract

Music tutor can provide practice advice for musical instrument leaners. The core is the combination of music theory knowledge, machine learning and chord recognition. Chord recognition is the basis of automatic music labeling, and plays an important role in music segmentation and audio matching. Aiming at the problem of low recognition rate of the same chord between different instruments, this paper uses an improved algorithm based on instantaneous frequency to extract Pitch Level Profile (PCP) features. Music instructor makes suggestions and plans for learners based on the mining data of chord recognition (accuracy rate, loudness difference, etc.) It can provide a more reasonable practice plan for beginners to make music teaching efficient.