Implementasi CNN untuk Klasifikasi Maqam Bacaan Alquran dengan Chroma Feature

Aryadi, M. Izdiar Alnafisi (2025) Implementasi CNN untuk Klasifikasi Maqam Bacaan Alquran dengan Chroma Feature. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5024211015-Undergraduate_Theses.pdf - Accepted Version Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Al-Qur'an sebagai kitab suci umat Islam memiliki seni baca (tilawah) yang dikenal dengan maqam, mencakup delapan irama utama seperti Bayati, Nahawand, Sika, Saba, Ajam, Kurdi, Hijaz, dan Rast. Klasifikasi maqam secara otomatis dapat ditingkatkan dengan ekstraksi fitur audio yang tepat. Penelitian ini bertujuan mengembangkan model klasifikasi maqam bacaan Al-Qur'an menggunakan Convolutional Neural Network (CNN) dengan memanfaatkan ekstraksi fitur Chroma Feature, yang secara teoritis lebih efektif menangkap karakter tonal maqam dibandingkan Mel-Frequency Cepstral Coefficients (MFCC). Dataset yang digunakan adalah Maqam478, terdiri atas delapan maqam dengan bacaan dari berbagai ayat Al-Qur'an. Hasil pengujian menunjukkan bahwa Chroma Feature mampu mengekstraksi pola irama dan perubahan nada yang unik pada setiap maqam, menghasilkan akurasi lebih tinggi daripada MFCC. Hasil nilai akurasi tertinggi yang didapatkan oleh Metode ekstraksi chroma feature adalah 97,27% yang merupakan hasil terbaik setelah dilakukan banyak oengujian dan dibandingkan dengan hasil akurasi model MFCC. Namun, variasi implementasi maqam oleh qari berbeda tetap menjadi tantangan. Penelitian ini menyimpulkan bahwa Chroma Feature layak dikembangkan lebih lanjut dengan dataset yang lebih beragam untuk meningkatkan generalisasi model.
===================================================================================================================================
The Qur'an, as the holy book of Islam, has a recitation art (tilawah) known as maqam, which includes eight main melodies such as Bayati, Nahawand, Sika, Saba, Ajam, Kurd, Hijaz, and Rast. Automatic maqam classification can be improved through proper audio feature extraction. This study aims to develop a classification model for Qur'anic recitation maqam using Convolutional Neural Network (CNN) by utilizing Chroma Feature extraction, which is theoretically more effective in capturing the tonal characteristics of maqam compared to Mel-Frequency Cepstral Coefficients (MFCC). The dataset used is Maqam478, consisting of eight maqams with recitations from various Qur'anic verses. The test results show that Chroma Feature can extract unique rhythmic patterns and pitch variations in each maqam, achieving higher accuracy than MFCC. The highest accuracy value obtained by the Chroma Feature extraction method is 97.27%, which is the best result after extensive testing and comparison with the MFCC model's accuracy. However, variations in maqam implementation by different reciters remain a challenge. This study concludes that Chroma Feature is worth further development with a more diverse dataset to improve model generalization.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Al-Quran, Maqam Bacaan, CNN, MFCC, Croma Feature, Sinyal Digital
Subjects:	T Technology > T Technology (General) > T57.6 Operations research--Mathematics. Goal programming T Technology > T Technology (General) > T57.74 Linear programming T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models. T Technology > T Technology (General) > T57.84 Heuristic algorithms. T Technology > T Technology (General) > T58.6 Management information systems T Technology > T Technology (General) > T58.62 Decision support systems
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Computer Engineering > 90243-(S1) Undergraduate Thesis
Depositing User:	M. Izdiar Alnafisi Aryadi
Date Deposited:	13 Jun 2025 08:59
Last Modified:	13 Jun 2025 08:59
URI:	http://repository.its.ac.id/id/eprint/119165

Actions (login required)

View Item