Abstract: Audio watermarking is a promising approach to copyright protection of audio data, especially music and songs. Several watermarking techniques have been developed and commercialized. The watermarks produced by those techniques can withstand a number of single attacks such as MPEG, resampling, filtering, and quantization. However, the watermarks are easily destroyed if subjected to chopping or multiple attacks. In this paper, we present a new audio watermarking technique that provides watermark robustness not only to single attacks but also to multiple attacks and chopping.