I. Introduction
The field of audio watermarking covers a wide variety of attacks, for example, clipping, collage, replacement by zeros, copy attacks, synchronization, content replacement, etc. [1] The growing number of attacks against watermarking systems has shown that there is the need for a way to describe their behaviour with the goal of designing schemes that are more robust against these attacks. A content replacement attack replaces a set of samples from an audio signal with another set of samples. In real scenarios, one of the more severe attacks against self-recovery schemes is the discordant size content replacement attack. In audio signals, this attack modifies the content, causing it to have a meaning that is different from the original. This impairs the integrity and authentication of digital signals.