In this age, information technology has grown significantly. Computer networks, which connect a device to others, have made it easier for people to transfer data than before. Moreover, smart devices have the capability of supporting this data transmission, including those in the cloud that may contain massive data. However, the security factor can be a severe issue if sensitive big data, such as military and medical data, do not have enough protection. Furthermore, an attacker may be able to disclose such data. Some algorithms have been introduced to solve that problem, one of which is the data hiding method. Nevertheless, some factors are still challenging, concerning the capacity of the secret data and the quality of the generated data, which are represented by bit and Peak Signal-to-Noise Ratio (PSNR), respectively. Besides, some techniques are not reversible, which means that they cannot reconstruct the carrier (cover). In this research, we investigate those problems by taking audio as the carrier. It is done by sampling the audio file before being interpolated to present spaces for accommodating the secret. Meanwhile, the secret is segmented before the embedding. Later, the embedded audio is smoothed according to the required level. The experimental result is obtained by using a public data set containing various audio genres and instruments, and 11 secret sizes, from 1 to 100 kb. It shows that the proposed method outperforms the others. This higher PSNR value means that the proposed method can generate more similar stego data; it also implies that at a certain quality level, the number of bits that can be hidden in the audio cover is higher than that of others.

在这个时代，信息技术得到了极大的发展。将设备连接到其他设备的计算机网络使人们比以前更容易传输数据。此外，智能设备具有支持这种数据传输的能力，包括可能包含大量数据的云中的那些设备。但是，如果敏感的大数据（例如军事和医疗数据）没有足够的保护，则安全因素可能成为严重问题。此外，攻击者可能能够披露此类数据。已经引入了一些算法来解决该问题，其中之一是数据隐藏方法。但是，关于秘密数据的容量和生成的数据的质量，仍然存在一些挑战，分别由比特和峰值信噪比（PSNR）表示。除了，有些技术是不可逆的，这意味着它们无法重建载波（封面）。在这项研究中，我们以音频为载体来研究这些问题。通过在对音频文件进行插值以呈现当前空间以容纳秘密之前对音频文件进行采样来完成此操作。同时，秘密在嵌入之前被分割。然后，根据所需的级别对嵌入式音频进行平滑处理。通过使用包含各种音频流派和乐器以及11种秘密大小（从1到100 kb）的公共数据集获得实验结果。结果表明，所提出的方法优于其他方法。较高的PSNR值意味着所提出的方法可以生成更多相似的隐身数据。这也意味着在一定质量级别上，音频封面中可以隐藏的位数要多于其他位数。