关键词: 音频水印, 幅度缩放, 小波变换, 抖动量化

Abstract: The robustness of audio watermarking is one of the research hotspots in watermarking.To improve the robustness against amplitude variation，a wavelet-based adaptive audio watermarking algorithm which can resist amplitude variation is proposed referring to the theory of digital image normalization.In the proposed algorithm，the audio file is decomposed adaptively and smoothed in space domain，and the watermark information is embedded into the low frequency coefficients of three levels wavelet transform with dither quantization.Experimental results show that the proposed algorithm is robust to amplitude variation，and it can resist common watermarking attacks such as additive noise，low pass filter，rep-sampling，and re-quantification.

Key words: audio watermarking, amplitude variation, wavelet transform, dither quantification