Muludi, Kurnia and Aristoteles, Aristoteles and Loupatty, SFB Frank Abe (2014) Chord Identification Using Pitch Class Profile Method with Fast Fourier Transform Feature Extraction. International Journal of computer science issues, 11 (3). ISSN 1694-0784


Download (715kB) | Preview


Abstract Recognizing a chord of a music is not an easy task to do, but it can be done easier using digital signal processing in computer system. This research implements pitch class profile (PCP) method in recognizing chord from audio signal. The audio signals are converted from time domain into frequency domain using Fast Fourier Transform (FFT). In this research, signals are processed using various frame blocking width in order to find out the most optimal width in frame blocking using Hamming window. This research implements chord recognition method in Python programming to process guitar chord recording. The purpose of this research is to find out the accuracy of chord recognition using PCP method. Keywords: Chord; Guitar; Pitch Class Profile; Fast Fourier Transform. 1. Introduction It is understood that chord is the most basic guitar playing. Chord playing is also the basic element in a song. To recognize chord is not an easy task to do, because chord doesn’t consist of single pitch. Moreover, a person with good musical ability is not surely capable to do it. A person with no musical experience needs long time effort to recognize a single pitch of music. But musical chord of audio signal can be recognized using signal processing technique with computer. Computer can recognize tone of a audio signal by measuring the fundamental frequency of a audio wave and match the frequency with the reference frequency. With signal processing, computer can recognize not only pitch but also other parameters of a signal those can be used to recognize a chord. Processing an audio signal takes a process called digital signal processing (DSP). DSP is a process that analyze input and output of a signal to know the certain physical system characteristic of a signal. Analyzing and synthesizing signal can be done easier at the frequency domain, because the determinant factor of signal is the frequency [12]. Therefore, the earliest stage that is necessary in DSP is feature extraction to convert signal from time domain into frequency domain. After that conversion, processing the signal can be done easier. In this research, a pitch identification system of audio signal has built to recognize the chord as the output. System was developed using fast fourier transform (FFT) as feature extraction technique and pitch class profile (PCP) as a method to identify the chord. This research shows the accuracy of chord recognition using PCP method.

Item Type: Article
Subjects: A General Works > AC Collections. Series. Collected works
Divisions: Fakultas Matematika dan Ilmu Pengetahuan Alam (FMIPA) > Prodi Ilmu Komputer
Depositing User: Fertika Dwi Yoswita
Date Deposited: 18 Jul 2016 07:50
Last Modified: 04 Nov 2016 02:52

Actions (login required)

View Item View Item