An Edge Device for Real-Time Fake and Tampered Audio Detection

The work presented a software-based framework for real-time fake and tampered audio detection using DSP techniques. The system analyzes short audio frames using ZCR variance, Auto-Correlation variance, and RMS variation to assess temporal activity, periodicity, and amplitude dynamics. The approach provides a lightweight and interpretable method for audio authenticity analysis; however, its reliability can be affected by noise, echo, compression, recording conditions, and advances in synthetic speech. Future work can improve robustness using additional features, diverse datasets, adaptive thresholds, and machine-learning or deep-learning models. Overall, the proposed system provides a foundation for real-time audio authenticity analysis.

[1]     R. Korycki, “Time and spectral analysis methods with machine learning for the authentication of digital audio recordings,” Forensic Science International, vol. 230, no. 1–3, pp. 117–126, 2013. [2]     DGMA, “Artificial Intelligence Approach To Improve Lifespan Of Wireless Sensor Network : A Survey “” International Conference on Innovation & Development of Future Smart Cities, pp. 1–9. [3]     J. Yamagishi et al., “ASVspoof 2021: Accelerating progress in spoofed and deepfake speech detection,” Proc. ASVspoof Challenge Workshop, pp. 47–54, 2021. [4]     J.-W. Jung et al., “AASIST: Audio Anti-Spoofing Using Integrated Spectro-Temporal Graph Attention Networks,” Proc. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 6367–6371, 2022, doi: 10.1109/ICASSP43922.2022.9747766. [5]     J. Yi, C. Wang, J. Tao, X. Zhang, C. Y. Zhang, and Y. Zhao, “Audio Deepfake Detection: A Survey,” IEEE Transactions on Dependable and Secure Computing, 2023. [6]     H. Tak, J. Patino, D. Delgado, M. Todisco, A. Nautsch, N. Evans, and A. Larcher, “End-to-End Anti-Spoofing with RawNet2,” Proc. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021. [7]     X. Wang et al., “ASVspoof 2019: A Large-Scale Public Database of Synthesized, Converted and Replayed Speech,” Computer Speech & Language, vol. 64, 2020. [8]     B. Ustubioglu, G. Tahaoglu, and G. Ulutas, “Detection of Audio Copy-Move-Forgery with Novel Feature Matching on Mel Spectrogram,” Expert Systems with Applications, vol. 213, 2023, Art. no. 118963, doi: 10.1016/j.eswa.2022.118963. [9]     H.-P. Hsu et al., “Detection of Audio Tampering Based on Electric Network Frequency Signal,” Sensors, vol. 23, no. 16, Art. no. 7029, 2023, doi: 10.3390/s23167029. [10]   Z. Almutairi and H. Elgibreen, “A Review of Modern Audio Deepfake Detection Methods: Challenges and Future Directions,” Algorithms, vol. 15, no. 5, Art. no. 155, 2022, doi: 10.3390/a15050155.