Comparative study of pitch detection algorithm to detect traditional Balinese music tones with various raw materials

Abstract

The characteristics of a tone are formed from the fundamental frequency values that make up the tone. A musical instrument’s tone has different fundamental values in building the timbre of the instrument’s tone. This study compares four PDA (Pitch Detection Algorithm) methods and tries to find the best algorithm for detecting Balinese instruments’ tone with different raw materials, namely leather, metal, and bamboo. The PDA method analyzed was the Autocorrelation, ZCR, HPS, and FFT methods. This study also explores the performance of each PDA (Pitch Detection Algorithm) approach in checking the pitch of traditional Balinese musical instruments and focuses on which of the four PDA (Pitch Detection Algorithm) methods detect Balinese musical instruments by comparing the results of the pitch analysis of each line with the original frequency. This study used 193 tone data consisting of Kendang to represent skin, Gangsa instruments to represent metal, Rindik, and Balinese flutes to represent bamboo material. Based on the test results, the PDA (Pitch Detection Algorithm) method that produces the closest value of the four musical instruments is HPS (Harmonic Product Spectrum) and FFT (Fast Fourier Transform) with an average difference of 4.81 Hz and 4.85 Hz.