Study of chord detection for music signals

Abstract

There are currently many pitch and chord detection applications, but most of them are only able to detect simple chords. As such, this project is a continuation from previous projects on chord detection for music signals, which aims to test the developed MATLAB algorithm with more complex chords consisting of more notes, in varying octaves and different inversions. Chords were played on the piano because of its vast chord range and octave variations, and analysis was done on the results for future improvements of the algorithm. The MATLAB algorithm was then compared against an Android application Chord Detector to determine its accuracy and efficiency. Other methods used to identify music chords without using algorithms or applications were also examined and evaluated based on their feasibility. It was concluded that the MATLAB algorithm was only able to detect chords with notes within the 3rd and 6th octaves of the piano. It was able to detect simple triads and moderately complex chords in root position, but was unable to detect complex chords comprising of 5 or more notes, as well as chords in the first and second inversions. It also had a higher accuracy rate compared to the Android application Chord Detector.