An Efficient Steganography Algorithm Based on QMDCT Coefficients in the Small Value Region of AAC
Zhenchao Shen, Fengyuan Nie, Bo Gao, Weiwei Liu · 2025
Advanced Audio Coding (AAC) is one of the most widely used audio compression coding standards on the Internet. Its high compression ratio, high sound quality, and strong robustness make it an ideal carrier for audio steganography. However, current adaptive steganography algorithms for AAC audio have high computational complexity, often leading to processing delays. To meet the requirements of lightweight algorithms for the development of streaming media, this paper proposes an efficient steganography algorithm based on quantized MDCT (QMDCT) coefficients in the small value region of AAC. The algorithm utilizes the characteristics of AAC encoding and the small value region coefficients, selecting QMDCT coefficients with lower energy in the small value region as the steganographic carrier. It also designs a lightweight distortion cost function that combines coefficient bit level, coefficient value, and Huffman coding codeword length. Through ternary matrix steganographic coding and the defined distortion cost function, the overall embedding distortion is minimized. Experimental results show that the algorithm has a high embedding capacity and good resistance to steganalysis.