Source Tracing of Audio Deepfake Systems

TL;DR

This paper presents a system that classifies specific techniques used in audio deepfake creation, helping to identify the distinct stages and methods involved in generating realistic fake audio samples.

Contribution

It introduces a novel classification system that detects various spoofing attributes across the entire audio deepfake generation pipeline, enhancing anti-spoofing capabilities.

Findings

System accurately classifies spoofing attributes on multiple datasets

Demonstrates robustness against different deepfake generation techniques

Improves understanding of deepfake generation stages

Abstract

Recent progress in generative AI technology has made audio deepfakes remarkably more realistic. While current research on anti-spoofing systems primarily focuses on assessing whether a given audio sample is fake or genuine, there has been limited attention on discerning the specific techniques to create the audio deepfakes. Algorithms commonly used in audio deepfake generation, like text-to-speech (TTS) and voice conversion (VC), undergo distinct stages including input processing, acoustic modeling, and waveform generation. In this work, we introduce a system designed to classify various spoofing attributes, capturing the distinctive features of individual modules throughout the entire generation pipeline. We evaluate our system on two datasets: the ASVspoof 2019 Logical Access and the Multi-Language Audio Anti-Spoofing Dataset (MLAAD). Results from both experiments demonstrate the robustness of the system to identify the different spoofing attributes of deepfake generation systems.