Abstract
In this thesis, we propose a digital watermarking scheme applied to digital audio signals to protect copyright. This watermark can be a 2D logo, a barcode image, or other information about the music. To satisfied two main features: inaudibility and robustness, we works on frequency domain. In this system, we do block DCT to original digital audio signal and embed watermarks into the coefficients. In order to get the embedded watermark in the correct position, we add a small signal in front of blocks for synchronization. We test this scheme under common audio editing processing, like re–sampling, re–resolution, and MP3 compression. Experiments show that we can extract watermark correctly in most situations, especially in MP3 compression.