Analisis Pengaruh Voice Changer Terhadap Pitch dan Forman Menggunakan Spektogram Dan Algoritma K-Nearest Neighbor

Risqullah, Moh Rifqy (2023) Analisis Pengaruh Voice Changer Terhadap Pitch dan Forman Menggunakan Spektogram Dan Algoritma K-Nearest Neighbor. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 01111940000028-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (9MB) | Request a copy

Abstract

Voice changer adalah salah satu bentuk modifikasi suara yang dilakukan dengan mengubah parameter fisis suara seperti pitch dan forman. Pada bidang forensik akustik, modifikasi tersebut dapat menjadi hambatan dalam proses membandingkan suara. Suara asli dan yang sudah dimodifikasi dapat dibandingkan secara visual melalui spektogram maupun secara analitik melalui parameter fisisnya. Penelitian ini bertujuan untuk melihat pengaruh voice changer terhadap parameter pitch (F0), forman 1 (F1), forman 2 (F2), dan forman 3 (F3) berdasarkan tampilan spektogram dan algoritma k-nearest neighbor (kNN). Algoritma kNN digunakan untuk mengklasifikasi 9 kata ucap dari 20 sampel suara laki-laki dan perempuan yang telah diberi voice changer berbeda yaitu: Female Voice (VCA), Male Voice (VCB), Telephone Effect (VCC), dan Radio Effect (VCD). Klasifikasi dilakukan untuk mengidentifikasi suara yang telah diberi voice changer berdasarkan jenis kelamin. Klasifikasi dilakukan melalui tiga parameter analisis yaitu F0-F1, F0-F2, dan F0-F3. Pemberian voice changer menghasilkan tampilan spektogram dan akurasi yang berbeda-beda pada domain frekuensi. Selain itu, beberapa sampel suara menunjukan hasil klasifikasi yang salah setelah diberi voice changer. Akurasi klasifikasi pada sampel suara laki-laki untuk VCA, VCC, dan VCD berturut-turut adalah 87%, 37%, dan 68%. Sedangkan akurasi pada sampel suara perempuan untuk VCB, VCC, dan VCD berturut-turut adalah 74%, 96%, dan 88%.
=================================================================================================================================
Voice changer is one of the forms of voice modification achieved by altering physical voice parameters such as pitch and formants. In the field of acoustic forensics, such modifications can pose challenges in the process of voice comparison. The original and modified voices can be compared visually through spectrograms as well as analytically through their physical parameters. This research aims to examine the effects of voice changers on pitch (F0), formant 1 (F1), formant 2 (F2), and formant 3 (F3) parameters based on spectrogram visualization and the k-nearest neighbor (kNN) algorithm. The kNN algorithm is used to classify 9 spoken words from 20 samples of male and female voices that have been given different voice changers, namely: Female Voice (VCA), Male Voice (VCB), Telephone Effect (VCC), and Radio Effect (VCD). The classification is conducted to identify the voice modified with different gender-based voice changers. The classification is performed using three parameter analyses: F0-F1, F0-F2, and F0-F3. The application of voice changers results in varying spectrogram visualizations and accuracy in the frequency domain. Additionally, some voice samples show incorrect classification results after being given a voice changer. The classification accuracies for male voice samples using VCA, VCC, and VCD are 87%, 37%, and 68%, respectively. Meanwhile, the accuracies for female voice samples using VCB, VCC, and VCD are 74%, 96%, and 88%, respectively.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Formant, K-Nearest Neighbor, Spectrogram, Pitch, Voice Changer, Forman, Spektogram
Subjects:	Q Science Q Science > QC Physics
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Physics > 45201-(S1) Undergraduate Thesis
Depositing User:	Moh Rifqy Risqullah
Date Deposited:	09 Nov 2023 07:12
Last Modified:	09 Nov 2023 07:12
URI:	http://repository.its.ac.id/id/eprint/103391

Actions (login required)

View Item