ANALISIS SISTEM PENGENAL PEMBICARA BERBASIS EKSTRAKSI CIRI FAST FOURIER TRANSFORM

Ardhita Pratama, Ryan (2023) ANALISIS SISTEM PENGENAL PEMBICARA BERBASIS EKSTRAKSI CIRI FAST FOURIER TRANSFORM. Thesis thesis, Institut Teknologi Dirgantara Adisutjipto.

	Text (ABSTRAK) 15010067 _ ABSTAK.pdf Download (103kB)
	Text (BAB I) 15010067 _ BAB 1.pdf Download (105kB)
	Text (DAFTAR PUSTAKA) 15010067 _ DAFTAR PUSTAKA .pdf Download (100kB)
	Text (FULL SKRIPSI) 15010067.pdf Restricted to Repository staff only Download (1MB)

Abstract

Sound is one way to communicate and express yourself. In this modern era, the need for systems and applications that are able to analyze and identify voice signals is even higher. The use of this application is also growing, ranging from learning tools to the security sector. Speech signal processing research in this Final Project uses the Fast Fourier Transform (FFT) algorithm. The Fast Fourier Transform algorithm divides the frequency per period. Therefore this algorithm can work well so as to produce accuracy quickly and efficiently. In this research, word recognition analysis will be carried out using the Fast Fourier Transform (FFT) method and Euclidean Distance classification. With the aim of introducing speakers with the accuracy of the training data and test data as evidenced in the form of the minimum Euclidean Distance results from all trials. The research results show that a speaker recognition system that can identify a speaker begins with the process of training data and test data. Speaker identification made using the Fast Fourier Transform (FFT) method and Euclidean Distance classification resulted in an average recognition accuracy of 85.7% which was carried out 7 times by testing the test data on 18 training data, with the best recognition accuracy of 100%.

Actions (login required)

View Item

Downloads