An Intelligent System for Audio Splicing Forgery Detection Using MFCC

One of the key forensic topics has been the detection of audio forgeries, mostly used as real evidence in court. As new methods of producing fake content emerge, digital audio/speech recordings are utilized as evidence that may be falsified and are capable of being detected if they have been. The transmission of digital audio/speech recording data over several media exposes the data to the risk of being attacked or tampered with. Several people misuse the altered audio and editing software, such as Adobe, Audition CC, etc., making it simple to manipulate the authentic audio, which results in audio forgeries. So, to overcome these scenarios, audio forgery detection method is deployed. Segments of the audio that have been altered are found using a forgery detection method. The system can verify if every part of the audio is from the claimed speaker. Nowadays, an audio recording does not incorporate digital marketing and signature mark for authentication due to the expensive process, in contrast to audio detection, which typically contains these elements for authenticity. Copy-move forgery and audio splicing are two of the most frequently employed forgery methods. Digital watermarking and signatures are commonly used methods for protecting and authenticating image and audio data. An audio tampering detection method for splicing is implemented in this paper. Mel frequency cepstral coefficient features are used in this paper to create the suggested approach. Combining recordings from two distinct speakers produces a dataset of altered audio to test the proposed method. A 90% accuracy rate was attained in this work.