Monophonic Guitar Notation Estimation Using YIN Pitch Detection Algorithm and Long Short-Term Memory
Sari
ABSTRAK
Proses transkripsi audio gitar ke dalam bentuk tablature masih menjadi tantangan bagi pemula karena adanya ambiguitas posisi nada pada fretboard, di mana satu nada yang sama dapat dimainkan pada kombinasi senar dan fret yang berbeda. Penelitian ini bertujuan untuk mengembangkan sistem estimasi tablature gitar monofonik yang efisien menggunakan kombinasi algoritma YIN dan model Long Short-Term Memory (LSTM). Algoritma YIN digunakan untuk mengekstraksi frekuensi fundamental (f0) dari sinyal audio WAV, yang kemudian divalidasi melalui serangkaian tahap post-processing untuk memastikan stabilitas nada. Model LSTM kemudian digunakan untuk memetakan urutan nada tersebut ke dalam representasi senar dan fret berdasarkan pola ergonomi jari. Hasil penelitian menunjukkan bahwa model LSTM yang dioptimasi menggunakan Optuna mencapai akurasi tertinggi sebesar 83,22%. Selain itu, penerapan parameter constraint pada model terbukti mampu menghasilkan rekomendasi posisi jari yang lebih ergonomis bagi pemain gitar dibandingkan model baseline. Sistem ini diimplementasikan dalam aplikasi berbasis web yang memungkinkan transkripsi audio ke tablature secara otomatis dengan beban komputasi yang rendah.
Kata kunci: Tablature Gitar, Monofonik, Algoritma YIN, LSTM, Transkripsi Musik
ABSTRACT
Transcribing guitar audio into tablature remains a significant challenge for beginners due to pitch ambiguity on the fretboard, where a single note can be played on multiple string and fret combinations. This research aims to develop an efficient monophonic guitar tablature estimation system using a combination of the YIN algorithm and Long Short-Term Memory (LSTM) models. The YIN algorithm is utilized to extract fundamental frequencies (f0) from WAV audio signals, which are further validated through various post-processing stages to ensure pitch stability. Subsequently, an LSTM model maps these pitch sequences into string and fret representations based on fingering patterns. The results indicate that the LSTM model optimized with Optuna achieved the highest accuracy of 83.22%. Furthermore, the implementation of parameter constraints in the model proved effective in generating more ergonomic fingering recommendations compared to the baseline model. This system is implemented as a web-based application, enabling automated audio-to-tablature transcription with low computational overhead.
Keywords: Guitar Tablature, Monophonic, YIN Algorithm, LSTM, Music Transcription.
Kata Kunci
Teks Lengkap:
PDFReferensi
M. Z. A. Rachmatullah, E. M. Yuniarno, and M. Attamimi, “Prediksi Akor Musik menggunakan Deep Learning berbasis Notasi Angka,” JURNAL TEKNIK ITS, vol. Vol. 10, no. No. 1, 2021.
S. Basini, G. N. Pardomuan, and M. S. Marlissa, “Pengenalan Dasar Alat Musik Gitar Untuk Siswa Kelas III Smp Negeri Borme Kabupaten Pegunungan Bintang Provinsi Papua,” Cantata Deo: Jurnal Musik dan Seni, vol. 1, no. 1, pp. 12–24, Apr. 2023, doi: 10.69748/jmcd.v1i1.7.
Y. Jadhav, A. Patel, R. H. Jhaveri, and R. Raut, “Transfer Learning for Audio Waveform to Guitar Chord Spectrograms Using the Convolution Neural Network,” Mobile Information Systems, vol. 2022, 2022, doi: 10.1155/2022/8544765.
A. Danika, J. Raharjo, and B. Hidayat, “Deteksi Suara Gitar Dengan Bahan Jenis Senar Berbeda Melalui Ciri Akustik Dengan Mel-Frequency Cepstral Coefficients (MFCC) Dan Support Vector Machine (SVM) Guitar String Detection Through Acoustic Characteristics Using Mel-Frequency Cepstral Coefficients (MFCC) And Support Vector Machine (SVM) Methods,” Bandung City, Dec. 2022.
P. Bonten and J. Kepler, “Guitar Tablature Detection using Recurrent Neural Networks,” Aug. 2022. [Online]. Available: www.jku.atDVR0093696
Y. Indrawaty, D. Rosmala, and A. M. Ramdhanial, “APLIKASI PEMBELAJARAN ALAT MUSIK GITAR MENGGUNAKAN MODEL SKENARIO MULTIMEDIA INTERAKTIF TIMELINE TREE,” Jurnal Informatika, vol. 4, no. 1, 2013.
D. Régnier, N. Martin, and L. Bigo, “Identification of rhythm guitar sections in symbolic tablatures,” Lille, Oct. 2021. [Online]. Available: https://hal.science/hal-03335822v1
B. Setiadi and E. B. Setiawan, “Aplikasi Penerjemah Tablatur Gitar Menggunakan Teknologi Augmented Reality Pada Platform Android,” ULTIMA InfoSys, vol. VII, no. 2, p. 86, 2016.
J. Casco-Rodriguez, “Rock Guitar Tablature Generation via Natural Language Processing,” Jan. 2023, [Online]. Available: http://arxiv.org/abs/2301.05295
Y. Afifah Noor, M. Prasetya Aji, and B. Astuti, “Analisis frekuensi Gitar Menggunakan Smartphone. Prosiding Seminar Nasional Pascasarjana UNNES,” Kota Semarang, 2020.
R. R. A. Shaleha, “Do Re Mi: Psikologi, Musik, dan Budaya,” Buletin Psikologi, vol. 27, no. 1, p. 43, Jun. 2019, doi: 10.22146/buletinpsikologi.37152.
K. Sulya, A. Wasika, K. Gede, D. Putra, D. Purnami, and S. Putri, “Klasifikasi Kunci Gitar Menggunakan Spectral Analysis dan K-Nearest Neighbor,” Kota Bali, Apr. 2020.
K. Muludi and A. Frank SFB Loupatty, “Chord Identification Using Pitch Class Profile Method with Fast Fourier Transform Feature Extraction,” May 2014. [Online]. Available: www.IJCSI.org
R. Solehudin, G. Hermawan, and S. Kom, “IMPLEMENTASI METODE MFCC (MEL FREQUENCY CEPSTRAL COEEFICIENT) DAN NAIVE BAYESIAN UNTUK KLASIFIKASI NADA DASAR GITAR,” Bandung City, Jan. 2019.
A. Wiggins and Y. Kim, “GUITAR TABLATURE ESTIMATION WITH A CONVOLUTIONAL NEURAL NETWORK,” Delft, Nov. 2019. [Online]. Available: https://github.com/andywiggins/tab-cnn
A. de Cheveigné and H. Kawahara, “YIN, a fundamental frequency estimator for speech and music,” J Acoust Soc Am, vol. 111, no. 4, pp. 1917–1930, Apr. 2002, doi: 10.1121/1.1458024.
L. Sukhostat and Y. Imamverdiyev, “A Comparative Analysis of Pitch Detection Methods Under the Influence of Different Noise Conditions,” Jul. 01, 2015, Mosby Inc. doi: 10.1016/j.jvoice.2014.09.016.
J. Kim, “Automatic Pitch Detection and Shifting of Musical Tones in Real Time,” 2014.
D. Jouvet and Y. Laprie, “Performance Analysis of Several Pitch Detection Algorithms on Simulated and Real Noisy Speech Data,” Aug. 2017. [Online]. Available: http://www.speech.kth.se/snack/
O. Babacan, T. Drugman, N. d’Alessandro, N. Henrich, and T. Dutoit, “A Comparative Study of Pitch Extraction Algorithms on a Large Variety of Singing Sounds,” Dec. 2019, [Online]. Available: http://arxiv.org/abs/1912.12609
D. Rahadian Fudholi et al., “The Application of LSTM in the AI-Based Enhancement of Classical Compositions,” vol. 7, no. 1, pp. 107–117, 2024, doi: 10.20895/INISTA.V7I1.1628.
N. Vatsya, A. Thipse, P. Dixit, R. Dafe, and K. Shejul, “Melody Generation using Deep Learning: Unleashing the Power of RNN and LSTM,” International Journal of Innovative Science and Research Technology (IJISRT), pp. 1713–1720, May 2024, doi: 10.38124/ijisrt/ijisrt24apr2001.
DOI: https://doi.org/10.26760/jrh.v10i1.51-66
Refbacks
- Saat ini tidak ada refbacks.
Alamat redaksi dan tata usaha:
Lembaga Penelitian dan Pengabdian Masyarakat Institut Teknologi Nasional
Fakultas, gedung 14 Lantai 3
Jl. PHH. Mustapa 23 Bandung 40124
Tlp. 022-7272215 Pes. 159, Fax. 022-7202892,
e-mail: hrekayasa@itenas.ac.id
Jurnal ini terlisensi oleh Creative Commons Attribution-ShareAlike 4.0 International License.


1.png)


