Blind Microphone Analysis and Stable Tone Phase Analysis for Audio Tampering Detection

In this paper, we present an audio tampering detection method, based on the combination of blind microphone analysis and phase analysis of stable tones, e.g., the electrical network frequency (ENF). The proposed algorithm uses phase analysis to detect segments that might have been tampered: These segments are further analyzed using a feature vector able to discriminate among different microphone types. Using this combined approach, it is possible to achieve a significantly lower false-positive rate and higher reliability as compared to standalone phase analysis.