A High-Payload Audio Watermarking Scheme for Real-Time Applications

Digital audio watermarking is mainly used for ownership protection of digital audio. In this research, an audio watermarking method is presented which can be adopted in real-time situations due to its high speed in both embedding and detecting stages. This approach has a higher payload in comparison with similar recent approaches. In the proposed algorithm, which uses a time-domain synchronization, to embed the watermark, the audio signal is first split into a number of blocks. Then the watermark is embedded into the FFT components of audio signal with greater magnitudes. Unlike common audio watermarking methods that use the 1-bit embedding approach, in the proposed scheme, a 2-bit embedding approach is presented which doubles the embedding payload compared to a number of similar works. The evaluation results on different audio signals demonstrate that the new scheme is faster and more transparent compared to its counterparts. The presented watermarking is found robust enough against the common attacks like additive noise, MP3 compression, low pass filtering, re-sampling and re-quantization. Using FPGA implementation/evaluation, the time complexities of the major parts of the proposed algorithm are reported.