A Robust Audio Watermarking Method Using Empirical Mode Decomposition

Abstract

EMD is used in this study to embed adaptive watermarks in audio recordings.   EMD seamlessly decomposes the audio signal into oscillatory Inherent Mode Functions (IMFs) from each subsequent frame.   The Intrinsic Mode Function (IMF) concludes with the time codes and watermark. The IMF's initial signal is exceedingly rare, rendering even a multitude of attacks largely ineffective on its quality.   Gaussian attack, pink noise, resampling, cropping, or filtering does not influence hidden marking.   Our watermarking method demonstrated superior efficacy compared to the recently disclosed tactics.