In this paper we intend to analyse, from a cryptographic perspective, a particular audio watermarking technique based on the Complex Spectral Phase Evolution (CSPE) algorithm. We will investigate the scheme mathematically (through the domains of practical cryptography and information security), theoretically (determining the characteristics of the scheme to evaluate its strengths and weaknesses) and experimentally (evaluating the results of the algorithm with different inputs, keys and variables to see if the technique has practical potential). We will investigate whether the technique replicates characteristics of Perfect Secrecy, a One-Time Pad, Random Oracle and the Avalanche effect. We will show how the CSPE audio watermarking scheme under investigation compares to current similar encryption techniques. Finally, we will outline possible areas to improve the technique and provide recommendations that will increase the security of the system so that it may be used in the fields of covert communication, steganography and cryptography.