Guitar Chord Recognition using MFCC Based Feature Extraction and Kaiser Windowing Linggo Sumarno
Electrical Engineering Study Program, Sanata Dharma University, Kampus III, Paingan, Maguwoharjo, Depok Sleman, Yogyakarta, Indonesia
lingsum[at]usd.ac.id
Abstract
Based on the previous studies of the guitar chord recognition systems, there is an indication that a study can still be carried out. In this case, a study to obtain a lower number of coefficients of feature extraction in a guitar chord recognition system can still be carried out. The purpose of this study is to obtain a lower number of coefficients of feature extraction in a guitar chord recognition system than the previous studies. In this study, the guitar chord recognition system uses MFCC (Mel Frequency Cepstral Coefficients) based feature extraction and Kaiser windowing. This study evaluated three parameters from the system, namely the lowest mel filter frequency and the number of mel filters in the mel filter bank, and also the shape factor of the Kaiser window. The results showed that by using only four coefficients of feature extraction, it could achieve an accuracy of up to 92.14%.
