In this paper, we present a novel audio watermark channel
model that characterizes more precisely the distortions in the
watermarked audio after mp3 attack. In previous work in the
field of watermarking, noise introduced by attacks was usually
modeled as AWGN. Spectrum analysis showed that a more
appropriate channel model, in the presence of MPEG coding,
is the Rayleigh fading channel model. Exactness of the
channel model and MPEG attack characterization is assessed
and the overall performance of the proposed model is
compared to the performance of the AWGN channel model.