Abstract: The music recommendation systems, currently in use rely upon ratings, likes and generic generalization of genres. This produces far from ideal recommendations and possibly constricts the exploration span of the listener, thus limiting one’s library to a set of popular artists and titles. In order to refine the recommendation system, scientific attributes of the track must be taken into account. These attributes can be represented in the form of vector parameters. These vector parameters can be meaningfully defined using a specially designed FFT algorithm and the derived data is sent to the main server which will serve as an open ended system. Thus resulting in a system which leads to an indirect recommendation performed by another user client.