Spread Embedding of Fragile Copyright Information to Protect Audio Signal

Nowadays, millions of gigabytes of new data are generated every second and a large part of it is multimedia data. The security of this large amount of data is an important problem as well as its transmission and storage. Data without proven authenticity should not be distributed or used without permission. Audio data, unlike other types of multimedia, is quite weak in terms of copyright protection. Industrial practices generally prioritize the quality of audio data over the security of copyright data, contrary to academic recommendations. In this study, a spread hash-supported copyright embedding algorithm is proposed to ensure the copyright protection of audio data. The proposed algorithm is tested on a total of 92.017 seconds of dataset consisting of 516 music files and the results are presented. The algorithm successfully performs copyright verification of any 2-millisecond fragment of the audio in any clipping attack. Despite the changes made to the entire audio data, a 6% Bit Error Rate and 0.9999 Normalized Correlation values are obtained. According to international standards for perceptual evaluation of audio quality, a score of ~-1.7 is obtained in the objective evaluation. All performance evaluations are presented with tables and graphs, and comparisons are made with similar models in the literature. This study is one of the first to use a spread-hashing technique for audio copyright protection and has shown high performance, especially against clipping attacks.