Detecting the Hidden: A Comprehensive Review of MP3 Steganalysis Techniques

MP3 is among the most widely adopted audio formats worldwide, frequently utilized in mobile and multimedia applications. Unfortunately, cybercriminals—such as hackers, hacktivists, and terrorist groups—have exploited this format by embedding steganographic content into MP3 files, rendering hidden information imperceptible to casual inspection. Steganography involves concealing secret messages within an audio cover file, enabling covert communication, whereas steganalysis refers to the detection and extraction of such hidden data. Traditional steganalysis methods primarily depend on statistical techniques to identify anomalies within the audio bitstream. Recent advancements, however, have introduced deep learning approaches capable of detecting subtle changes and complex patterns in audio signals. This review examines key aspects of MP3 steganalysis, including dataset selection, feature extraction, and detection methodologies, drawing on both statistical and deep learning-based frameworks. Effective analysis hinges on extracting spectral, temporal, and statistical features while addressing challenges such as compression artifacts, variable encoding parameters, and fluctuating bitrates. Despite considerable progress, significant research gaps remain—particularly in detecting low-capacity embeddings, adapting to variable sample lengths, and ensuring model generalization across diverse MP3 encoding formats. Notably, this study is grounded in a comprehensive analysis of high-quality research articles published in reputable databases such as IEEE Xplore, Scopus, and Web of Science, ensuring the reliability and relevance of the findings. To further advance MP3 steganalysis, future work must focus on curating diverse datasets, refining feature extraction strategies, and developing hybrid detection models to enhance forensic robustness.