anushadatar / chordrecognition

Final Project for Olin College of Engineering Linearity One. Discrete Fourier Transform/Fast Fourier Transform Study and Application to Note and Chord Recognition.

Chord Recognition: The Fourier Transform and Automated Classification

Olin College Linearity I SP18 Final Project, Alison Palmer and Anusha Datar.

This project was mainly an exploration into the mathematics and physical concepts associated with Fourier analysis and its application to note and chord recognition. Along with a paper, we wrote and documented executable note and chord recognition programs that print the identity of a tone contained in a .wav file passed to the program. Models trained based on kaggle datasets cited in the /data directory.

chordrecognition's Issues

It can't recognize the note below a2.

Even if use the data in the music_notes folder. When it meets the note below a2, such as g2 e2 f2 c2, it will give the wrong answer. I guess the possible reason is that the Blackman window filter this frequency.