Klasifikasi Suara Aransemen Alat Musik Tiup Menggunakan Metode Reccurent Neural Network (RNN)

Putri, Dwinggrit Oktaviani (2025) Klasifikasi Suara Aransemen Alat Musik Tiup Menggunakan Metode Reccurent Neural Network (RNN). Undergraduate thesis, UPN Veteran Jawa Timur.

	Text Cover.pdf Download (1MB)
	Text (Bab 1) Bab 1.pdf Download (135kB)
	Text (Bab 2) Bab 2.pdf Restricted to Repository staff only until 18 September 2027. Download (382kB)
	Text (Bab 3) Bab 3.pdf Restricted to Repository staff only until 18 September 2027. Download (282kB)
	Text (Bab 4) Bab 4.pdf Restricted to Repository staff only until 18 September 2027. Download (534kB)
	Text (Bab 5) bab 5.pdf Download (110kB)
	Text (Daftar Pustaka) Daftar Pustaka.pdf Download (90kB)
	Text (Lampiran) Lampiran.pdf Restricted to Repository staff only Download (665kB)

Abstract

Wind instruments such as saxophone, clarinet, trumpet, and others have distinct acoustic characteristics, including frequency, amplitude, and waveform, which serve as unique identifiers for each instrument. In the digital era, the classification of musical instruments for song production or arrangement is still largely performed manually. Unlike previous studies that generally focused on musical instrument classification in a broad sense, this research introduces a more specific approach targeting four types of wind instruments, aiming to improve classification accuracy and reduce misclassification among instruments with similar timbres. This study aims to develop a wind instrument classification system using the Recurrent Neural Network (RNN) method. The classification process is based on extracted acoustic features such as Mel-Frequency Cepstral Coefficients (MFCC) and spectrograms or sonograms. Following feature extraction, recurrent cells in the form of LSTM and GRU are employed to capture temporal patterns from sequential data, particularly audio signals. The dataset consists of 1,200 audio files (.wav) representing four wind instruments (trumpet, baritone, mellophone, and tuba) with 300 audio samples for each instrument. The experimental results show that the LSTM model achieved an accuracy of 94%, while the GRU model reached an accuracy of 98%. These findings highlight the effectiveness of RNN, specifically the LSTM and GRU architectures, in classifying wind instruments with high accuracy. This study is expected to serve as a foundation for further development in the field of digital music, particularly in supporting automatic arrangement processes.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Prasetya, Dwi Arman	NIDN0005128001	arman.prasetya.sada@upnjatim.ac.id
Thesis advisor	Saputra, Wahyu Syaifullah Jauharis	NIDN0725088601	wahyu.s.j.saputra.if@upnjatim.ac.id

Subjects:

Q Science > Q Science (General)

Divisions:

Faculty of Computer Science > Departemen of Data Science

Depositing User:

Dwinggrit Oktaviani Putri

Date Deposited:

19 Sep 2025 03:17

Last Modified:

19 Sep 2025 03:17

URI:

https://repository.upnjatim.ac.id/id/eprint/43744