Evaluating Fake Music Detection Performance Under Audio Augmentations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sroka, Tomasz, Wężowicz, Tomasz, Sidorczuk, Dominik, Modrzejewski, Mateusz
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909688989220864
author Sroka, Tomasz
Wężowicz, Tomasz
Sidorczuk, Dominik
Modrzejewski, Mateusz
author_facet Sroka, Tomasz
Wężowicz, Tomasz
Sidorczuk, Dominik
Modrzejewski, Mateusz
contents With the rapid advancement of generative audio models, distinguishing between human-composed and generated music is becoming increasingly challenging. As a response, models for detecting fake music have been proposed. In this work, we explore the robustness of such systems under audio augmentations. To evaluate model generalization, we constructed a dataset consisting of both real and synthetic music generated using several systems. We then apply a range of audio transformations and analyze how they affect classification accuracy. We test the performance of a recent state-of-the-art musical deepfake detection model in the presence of audio augmentations. The performance of the model decreases significantly even with the introduction of light augmentations.
format Preprint
id arxiv_https___arxiv_org_abs_2507_10447
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Evaluating Fake Music Detection Performance Under Audio Augmentations
Sroka, Tomasz
Wężowicz, Tomasz
Sidorczuk, Dominik
Modrzejewski, Mateusz
Sound
Artificial Intelligence
Machine Learning
Audio and Speech Processing
With the rapid advancement of generative audio models, distinguishing between human-composed and generated music is becoming increasingly challenging. As a response, models for detecting fake music have been proposed. In this work, we explore the robustness of such systems under audio augmentations. To evaluate model generalization, we constructed a dataset consisting of both real and synthetic music generated using several systems. We then apply a range of audio transformations and analyze how they affect classification accuracy. We test the performance of a recent state-of-the-art musical deepfake detection model in the presence of audio augmentations. The performance of the model decreases significantly even with the introduction of light augmentations.
title Evaluating Fake Music Detection Performance Under Audio Augmentations
topic Sound
Artificial Intelligence
Machine Learning
Audio and Speech Processing
url https://arxiv.org/abs/2507.10447