Benchmark RadarMADBench
MADBench is a benchmark for modality-aware audio deepfake detection, treating speech and environmental audio as distinct components. It evaluates detectors across independently manipulated forgery sources.
- Released
- 2026-08-10
- Readiness
- Paper only
- Primary field
- Robotics & Autonomous Systems
Why it matters
It addresses the gap in audio deepfake detection by distinguishing speech and background audio, which have different generative mechanisms and artifact profiles, enabling component-aware evaluation.
Motivation
Recent advances in speech synthesis and audio generation have made high-fidelity acoustic forgery low-cost and difficult to attribute, enabling a realistic attack scenario in which speech and background audio are independently manipulated over otherwise authentic video.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.