BCH encoded robust and blind audio watermarking with tamper detection using hash

International Federation of Phonographic Industry (IFPI) has laid requirements for digital audio watermarking (DAW) schemes with regard to robustness, imperceptibility and payload. In addition, DAW schemes have to offer audio integrity. Many of the schemes fails to provide both copyright protection and tamper detection. Moreover, tamper detection of audio is not evaluated in quantitative manner. Therefore, a robust and blind DAW scheme has been developed in transform domain to suffice this requirement. Further, an objective performance metric, detection accuracy, is proposed to quantify the performance of tamper detection. In the proposed work, audio is fragmented and transform coefficients are obtained using Wavelet and Cosine Transforms. Watermark image is pre-processed using logistic map and BCH code. This pre-processed image is inserted into singular values of audio signal coefficients through quantization index modulation. Hash is generated with SHA-512 algorithm and is inserted in watermarked audio frames. Hash recovery at the receiver allows to identify any tampering involved in audio viz., deletion, copy-move and substitute attacks. Singular vectors are not required to extract the watermark, thus this method is free from false positive problem and can provide copyright protection. Experimentation is carried out on five different classes of audios which substantiates that robustness, imperceptibility and payload results are acquiescent to IFPI standard. Further, comparative analysis indicates that developed scheme is better when compared with other state-of art-schemes. Consequently, this DAW scheme helps in forensic examination of audio recording for authentication purpose.