AUDIO WATER-MARKING METHOD USING ANALYSIS CHARACTERISTIC OF WAVELET CONVERSION

摘要

PURPOSE: An audio water-marking method using an analysis characteristic of a wavelet conversion is provided, which generates a water-mark which has a noise and satisfies a strong characteristic to an attack using a wavelet conversion and a mentality sound model and prevents an unlawful copy of an audio data. CONSTITUTION: An audio water-marking method comprises the steps of: converting an audio signal into a wavelet using a wavelet conversion method and analyzing the converted value to an approximation value and a detail value; compressing an original signal using a CODEC and restoring the compressed signal, performing a wavelet conversion to compute S64; computing a difference value of the original value and S64 approximation value; multiply a value of an approximation value of the original value and a PN-sequence value and multiplying each sample by an other scale vector to generate a random water-mark; adding the difference value to the generated water-mark to generate a water-mark; performing a wavelet conversion and extracting a water-mark; and testing a correlation with an original water-mark value and verifying an ownership of a water-mark. Thereby, it is possible to prevent an unlawful copy.