ANALISIS SISTEM PENGENAL NADA BERBASIS FAST FOURIER TRANSFORM (FFT) DAN DYNAMIC TIME WARPING (DTW) UNTUK MENGENALI NADA PIANO

Bondan NurFauzi Rambe, Ahmad (2022) ANALISIS SISTEM PENGENAL NADA BERBASIS FAST FOURIER TRANSFORM (FFT) DAN DYNAMIC TIME WARPING (DTW) UNTUK MENGENALI NADA PIANO. Skripsi thesis, Institut Teknologi Dirgantara Adisutjipto.

	Text (ABSTRAK) 18010037_ABSTRAK.pdf Download (21kB)
	Text (BAB I) 18010037_BAB I.pdf Download (79kB)
	Text (DAFTAR PUSTAKA) 18010037_DAFTAR PUSTAKA.pdf Download (136kB)
	Text (FULL SKRIPSI) 18010037.pdf Restricted to Repository staff only Download (2MB)

Abstract

Musical instruments are anything that can be used to create notes and rhythms. The piano has 7 basic notes/chords consisting of A, B, C, D, E, F and G. To enjoy the strains of music that is heard, it is necessary to understand the tones created by the musical instrument itself. Like the development of music technology today which is very sophisticated where a music player can adjust the chords of a musical instrument through the help of applications on Android, for example, setting chords on guitar, piano, and biola. In this final project, the author designed a tone recognition system based on piano tone using matlab software with dynamic time warping classification method. Dynamic time warping is widely used for voice recognition, handwriting and signature recognition, gesture and music processing. The initial stage in piano tone recognition is to create training data and test data that go through the same process stages, namely voice recording, preprocessing, and FFT feature extraction. Then the values obtained are processed in the dynamic time warping classification by getting the smallest loca distance value from the comparison between training data and test data. The total voice data used is 56 voice signals with 35 used as training data and 21 as test data. Each note is recorded 8 times with 5 tones as training data and 3 tones as test data. The notes recorded are A, B, C, D, E, F, and G tones. From the test results using 3 test data compared to 20 training data, the highest success rate is 100% for the recognition of B, C, D tones, and for the recognition of A tones, the success rate is 66.67%. So from the overall accuracy value obtained for all piano tone recognition, the average value of tone recognition accuracy is 92%.

Actions (login required)

View Item

Downloads