Controlling Quality and Payload in Reversible Data Hiding Based on Modified Error Expansion for Segmental Audio Waveforms

Reversible data hiding is a technique where hidden data are embedded in host data such that the consistency of the host is perfectly preserved and its data are restored during extraction of the hidden data. In this paper, improvements are proposed for a linear prediction technique for reversible data hiding of audio waveforms. The proposed modified error-expansion method is able to control the sound quality degradation of the stego signal by varying the payload size of segmented audio waveforms. Detection of tampered stego frames is possible through the modified technique, while the hosts of uncorrupted stego signals can be perfectly recovered. Additionally, concealment of a payload within the stego data is better than by the previous method.