Digital watermarking based on process of speech production

摘要: A speech production procedure can be divided into three parts, namely the glottal source, articulation and radiation, respectively. We propose a watermarking method for by manipulating in the process of production. apply our to CS-ACELP(G.729 standard), which is ITU-T approved recommendation. It provides low bit rate 8 kb/s coding algorithm with wire/line quality. The watermarked vocal tract model expressed codebooks made LSP(Line Spectrum Pair) parameters. The codebook vectors replace some extracted LSP. Speech synthesized using replaced generate a couple unique modify LSP spectrum envelope. Shortening width of LSPs creates one watermarked codebook, and second codebook created stretching of both sides each formant. There are ten dimensions voice frame CS-ACELP decoder. In the detecting process, weighted Euclidean distance(WED) between the extracted will calculated. Whether watermark embedded judged utilizing calculated WED. Evaluation tests on detection accuracy discussed simulation results.