RTCFake: Speech Deepfake Detection in Real-Time Communication

2026-04-26

AI Summary

This study addresses the limited generalization of existing deepfake speech detection methods in real-world real-time communication (RTC) scenarios, where complex distortions introduced by codec compression and unseen speech enhancement techniques—such as noise suppression—significantly degrade performance. To bridge this gap, the authors construct RTCFake, the first large-scale deepfake speech dataset tailored for RTC environments, comprising approximately 600 hours of audio captured via mainstream social and conferencing platforms to precisely align offline-generated and online-transmitted samples. They further propose a phoneme-guided consistency learning (PCL) strategy to encourage the model to learn platform-invariant semantic structural representations. Experiments demonstrate that the proposed approach substantially outperforms current methods on evaluation sets involving unseen RTC platforms and challenging acoustic conditions, markedly enhancing both robustness and cross-platform generalization of deepfake detectors.

Technology Category

• Natural Language Processing: Speech
• Machine Learning: Life-Long and Continual Learning
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Security and Privacy: Large-scale security measurements
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

With the rapid advancement of speech generation technologies, the threat posed by speech deepfakes in real-time communication (RTC) scenarios has intensified. However, existing detection studies mainly focus on offline simulations and struggle to cope with the complex distortions introduced during RTC transmission, including unknown speech enhancement processes (e.g., noise suppression) and codec compression. To address this challenge, we present the first large-scale speech deepfake dataset tailored for RTC scenarios, termed \textit{RTCFake}, totaling approximately 600 hours. The dataset is constructed by transmitting speech through multiple mainstream social media and conferencing platforms (e.g., Zoom), enabling precise pairing between offline and online speech. In addition, we propose a phoneme-guided consistency learning (PCL) strategy that enforces models to learn platform-invariant semantic structural representations. In this paper, the RTCFake dataset is divided into training, development, and evaluation sets. The evaluation set further includes both unseen RTC platforms and unseen complex noise conditions, thereby providing a more realistic and challenging evaluation benchmark for speech deepfake detection. Furthermore, the proposed PCL strategy achieves significant improvements in both cross-platform generalization and noise robustness, offering an effective and generalizable modeling paradigm. The \textit{RTCFake} dataset is provided in the {https://huggingface.co/datasets/JunXueTech/RTCFake}.

Problem

• speech deepfake detection
• real-time communication
• codec compression
• speech enhancement
• cross-platform generalization

Innovation

• RTCFake
• speech deepfake detection
• real-time communication
• phoneme-guided consistency learning
• cross-platform generalization