On the automated recognition of seriously distorted musical recordings

In this paper, a new methodology is presented for the automated recognition -identification of musical recordings that have suffered from a high degree of playing speed and frequency band distortion. The procedure of recogniti on is essentially based on the comparison between an unknown musical record ing and a set of model ones, according to some predefined specific characte ristics of the signals. In order to extract these characteristics from a mu sical recording, novel feature extraction algorithms are employed. This pro cedure is applied to the whole set of model musical recordings, thus creati ng a model characteristic database. Each time we want an unknown musical re cording to be identified, the same procedure is applied to it, and subseque ntly, the derived characteristics are compared with the database contents v ia an introduced set of criteria. The proposed methodology led to the devel opment of a system whose performance was extensively tested with various ty pes of broadcasted musical recordings. The system performed successful reco gnition for the 94% of the tested recordings. It should be noted that the p resented system is parallelizable and can operate in real time.