Segmentation embedding method with modified interpolation for increasing the capacity of adaptable and reversible audio data hiding

Abstract

The development of digital technology brings many advantages to human life. However, it also comes with the risk of making data protection required. This protection can come in many forms, one of which is the data hiding method, often called steganography. It works by embedding secret data into media data, such as images, audio, and video, to avoid suspicion caused by their existence. Nevertheless, the embedding capacity and stego-data quality problems have become challenging. This research proposed new audio steganography, which utilizes modified linear interpolation values to give flexibility to the embedding capacity. This proposed method also adds bit thresholds to distribute the embedding capacity as evenly as possible to maintain its similarity level (quality) as high as possible by shifting the embedded sample value towards the original value of stego-data produced. The experimental results show that this method can obtain an average similarity level of 100.55 dB on 0.75 bit per sample audio data and 87.68 dB on 2.267 bit per sample audio data. The stability of stego-data quality is also measured by considering the best and minimum quality measurements with a difference of less than 1 dB on embedding secret data below 1.52 bit per sample.