Steganographic model to conceal the secret data in audio files utilizing a fourfold paradigm: Interpolation, multi-layering, optimized sample space, and smoothing

Abstract

Audio steganography techniques are evaluated using performance metrics such as payload capacity, stego audio file quality, and security (imperceptibility of hidden data). Increasing payload capacity often reduces stego image quality and security. However, maintaining high-quality stego audio with high imperceptibility requires compromising payload capacity. Therefore, achieving a balanced trade-off between payload capacity and stego audio quality has become essential but increasingly challenging. Current methods often lack adaptiveness, potentially compromising both embedding capacity and stego audio quality. To address these limitations, this study offers valuable insights to guide researchers in developing high-performing audio steganography models. The proposed method seeks to improve stego audio quality by implementing a smoothing-based technique and optimizing the sample space through linear interpolation, followed by a multi-layering process. Comprehensive experiments are conducted on common audio datasets to benchmark the method against state-of-the-art techniques. The results demonstrate peak signal-to-noise ratio (PSNR) values ranging from 105.14 to 126.34 dB, indicating substantial fidelity preservation compared to existing methods with PSNR values ranging from 120.55 decibel (dB) for 1 kb to 100.57 dB for 100 kb. These findings confirm the effectiveness of the model in maintaining audio quality while concealing data. Moreover, this study recommends that practitioners in steganography with audio data prioritize data concealment algorithms that enhance security over stego audio quality.