With advanced audio manipulation raising authenticity concerns, this study introduces a method to detect copy-move forgeries, where segments are duplicated and reinserted. It combines dominant resonance frequency analysis and dynamic time warping for high-precision detection, surpassing correlation-based methods. The approach segments audio, applies energy thresholds, and analyzes formant contours to track manipulations. Dynamic time warping provides a key advantage by dynamically aligning speech segments and effectively handling nonuniform distortions and pitch variations that traditional methods fail to capture. This capability significantly enhances its ability to detect temporally inconsistent duplications. Extensive evaluations on the Defense Advanced Research Projects–Texas Instruments/Massachusetts Institute of Technology Acoustic-Phonetic Continuous Speech Corpus (TIMIT; English) and the Indian Institute of Technology Kharagpur Simulated Emotion Hindi Speech Corpus (IITKGPSEHSC; Hindi) data sets demonstrate this method’s exceptional accuracy and cross-lingual adaptability, reliably detecting subtle copy-move forgeries with superior precision even in complex voiced segments. This method’s practical applications extend to forensic investigations, combating misinformation and enhancing cybersecurity by ensuring audio content integrity.
(See document for exact affiliation information.)
https://aes2.org/publications/elibrary-page/?id=22925
(668KB)
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member Join the AES. If you need to check your member status, login to the Member Portal.
