Identical Instruments Detection using Least Mean Square Based Adaptive Filter

Abstract

Gamelan is one of Indonesian traditional music instrument. It have variations in terms of fundamental frequency, amplitude, and signal envelope, due to its handmade construction and playing style. When playing gamelan instruments, many identical music is played simultaneously. Detection of identical instruments are needed to indicate the changes of Gamelan features. We can find how many identical instruments are played at the same time or separated by very short time. We applied Least Mean Square (LMS) adaptive filter to detect identical instruments. The reference signal can have a fundamental frequency shift from the original signal to 10 Hz. The synthetic saron signal is used as reference signal. The performance can separate the identical instruments in 0.03 ms.