Automatic Musical Genre Classification Of Audio Signals-英文文献.pdf

Automatic Musical Genre Classification Of Audio Signals George Tzanetakis Georg Essl Perry Cook Computer Science Department Computer Science Dep. Computer Science and Music Dep. 35 Olden Street 35 Olden Street 35 Olden Street Princeton NJ 08544 Princeton NJ 08544 Princeton NJ 08544 +1 609 258 5030 +1 609 258 5030 +1 609 258 5030 gtzan@ gessl@ prc@ ABSTRACT Musical genres are categorical descriptions that are used to In this work, algorithms for automatic genre classification are describe music. They are commonly used to structure the explored. A set of features for representing the music surface and increasing amounts of music available in digital form on the rhythmic structure of audio signals is proposed. The performance Web and are important for music information retrieval. of this feature set is evaluated by training statistical pattern Genre categorization for audio has traditionally been recognition classifiers using audio collections collected from performed manually. A particular musical genre is compact disks, radio and the web. Audio signals can be automatically classified using a hierarchy of genres that can be characterized