ABSTRACT
The automated discovery of patterns within music is an important problem within computational music analysis. Patterns emerge from repetitions within the music itself, and these patterns and repetitions are generally a good indication to musical structure. Repetitions can range in size from simple repeating note sequences, to recurring sections or phrases within a musical piece. In much the same way, structure can be defined at differing levels of granularity based on these patterns and repetitions; from high level song form down to the single note level, if required. In addition, each of these structures in turn can be interrelated and higher level structures can contain complex organization within themselves. In the audio compression research community there has being a common consensus that there is no apparent repetitiveness and approximate repetitiveness in audio. Some agree however that repetitiveness occurs but shifted in phase, difficult to recognize in the time occurs but shifted in phase, so difficult to recognize in the time domain [1, 2]. We however have defined earlier what we believe the pattern to be with regards music therefore the trust of this paper is that pattern matching audio (uncompressed and compressed) is possible.
