Template-based Chord Recognition : Influence of the Chord Types.

This paper describes a fast and efficient template-based chord recognition method. We introduce three chord models taking into account one or more harmonics for the notes of the chord. The use of pre-determined chord models enables to consider several types of chords (major, minor, dominant seventh, minor seventh, augmented, diminished...). After extracting a chromagram from the signal, the detected chord over a frame is the one minimizing a measure offit between the chromagram frame and the chord templates. Several popular measures in the probability and signal processing field are considered for our task. In order to take into account the time persistence, we perform a post-processing filtering over the recognition criteria. The transcription tool is evaluated on the 13 Beatles albums with different chord types and compared to state-of-theart chord recognition methods. We particularly focus on the influence of the chord types considered over the performances of the system. Experimental results show that our method outperforms the state-of-the-art and more importantly is less computationally demanding than the other evaluated systems.