Perancangan sistem pengukur similaritas musik = Design of music similarity measuring system

Sumaki, Kezia Stacy (2022) Perancangan sistem pengukur similaritas musik = Design of music similarity measuring system. Bachelor thesis, Universitas Pelita Harapan.

[img]
Preview
Text (Title)
Title.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (20kB) | Preview
[img]
Preview
Text (Abstract)
Abstract.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (111kB) | Preview
[img]
Preview
Text (ToC)
ToC.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (265kB) | Preview
[img]
Preview
Text (Chapter1)
Chapter1.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (52kB) | Preview
[img] Text (Chapter2)
Chapter2.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (486kB)
[img] Text (Chapter3)
Chapter3.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (140kB)
[img] Text (Chapter4)
Chapter4.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (618kB)
[img] Text (Chapter5)
Chapter5.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (2MB)
[img] Text (Chapter6)
Chapter6.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (114kB)
[img]
Preview
Text (Bibliography)
Bibliography.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (175kB) | Preview
[img] Text (Appendices)
Appendices.pdf
Restricted to Repository staff only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (3MB)

Abstract

Pesatnya perkembangan industri musik membuat kombinasi elemen musik menjadi lebih susah untuk diracik dan mengakibatkan adanya kemiripan antara satu musik dengan musik lainnya yang dapat berakhir dalam kontroversi. Penelitian ini bertujuan untuk membuat sistem pengukur similaritas antara dua sampel musik. Similaritas diukur dengan mencari korelasi rata-rata dari data kekuatan sinyal di banyak titik waktu pada spektrum frekuensi. Data ini, yang divisualisasikan dengan spectrogram, dapat memperlihatkan pola dan keunikan musikalisasi setiap sampel musik. Terdapat beberapa skema percobaan yang bertujuan untuk menguji kinerja sistem pengukur similaritas lagu dan untuk melihat pengaruh suatu faktor dalam pengukuran similaritas. Sistem dapat mengidentifikasi kemiripan warna nada dari tipe maupun jenis alat musik yang berbeda dengan rentang similaritas 0,68-0,87 untuk sistem pengembangan pertama dan 0,63-0,91 untuk sistem pengembangan kedua yang merupakan sistem pengukur similaritas musik. Pada percobaan selanjutnya, didapatkan bahwa perbedaan nada akan menghasilkan nilai similaritas yang berkisar di 0,01-0,21. Perbedaan durasi penekanan nada dan posisi nada tidak terlalu berpengaruh karena masih memberikan nilai similaritas pada rentang 0,81-0,93. Selanjutnya, pemotongan terhadap audio asli akan sedikit menurunkan hasil similaritas menjadi sekitar 0,97. Perbedaan kompleksitas atau perbedaan banyaknya nada pada satu titik memberikan nilai similaritas pada rentang 0,05-0,09. Pada percobaan lainnya, perbedaan genre lagu akan menghasilkan nilai similaritas yang berkisar di 0,18-0,38. Namun, perbedaan genre yang diikuti dengan perbedaan stuktur musik akan menghasilkan nilai similaritas lebih rendah yang berkisar di 0,12-0,17. Pada skema terakhir, sistem pengukur similaritas musik dapat digunakan sebagai music search engine. Sistem dapat membuat audio rekaman memberikan nilai similaritas 0,19-0,45 terhadap lagu aslinya dan 0,06-0,23 terhadap lagu yang mirip dengan lagu aslinya. Dengan demikian, faktor yang paling berpengaruh dalam penurunan nilai similaritas adalah perbedaan nada. / The rapid development of the music industry makes it harder to produce a new pleasing-to-ear musical components combination. At times, there would be a part of a music sounding similar to other music, resulting in a new controversy. This study aims to create a system to calculate the similarity between two music samples. Similarity is calculated by finding the average correlation of the signal strength at many time points on the frequency spectrum. This data, which is visualized by spectrogram, can show the musical pattern and uniqueness of each music sample. There are several experimental schemes that aim to test the performance of the system and to observe the effect of a factor in similarity measurement. The system can identify the timbre similarity of different type or kind of musical instruments with a similarity range of 0.68-0.87 for the first development system and 0.63-0.91 for the second development. In the next experiment, it was found that the difference in musical note will produce similarity values ranging from 0.01 to 0.21. The difference duration of note sound and its position does not have much effect because it still gives similarity values in the range of 0.81-0.93. Furthermore, truncation of the original audio will slightly lower the similarity results to around 0.97. The difference in complexity or the difference in the number of notes at one time point gives a similarity value in the range of 0.05-0.09. In other experiments, different song genres will produce similarity values ranging from 0.18-0.38. However, differences in genre followed by differences in musical structure will result in lower similarity values ranging from 0.12 to 0.17. In the last scheme, the system can be used as a music search engine. The system can make an audio recording sample to give similarity values of 0.19-0.45 if compared to its original song and 0.6-0.23 if compared to songs that are similar to its original song. Thus, the most significant factor in decreasing the similarity value is the difference in musical note.

Item Type: Thesis (Bachelor)
Creators:
CreatorsNIMEmail
Sumaki, Kezia StacyNIM01032180011keziasumaki@gmail.com
Contributors:
ContributionContributorsNIDN/NIDKEmail
Thesis advisorMartoyo, IhanNIDN0318057301ihan.martoyo@uph.edu
Uncontrolled Keywords: Audio processing; Frequency spectrum; Musical similarity
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions: University Subject > Current > Faculty/School - UPH Karawaci > Faculty of Science and Technology > Electrical Engineering
Current > Faculty/School - UPH Karawaci > Faculty of Science and Technology > Electrical Engineering
Depositing User: Users 9171 not found.
Date Deposited: 21 Feb 2022 03:12
Last Modified: 21 Feb 2022 03:12
URI: http://repository.uph.edu/id/eprint/46508

Actions (login required)

View Item View Item