Skip to Main Content
Although drum loops are widely present in many audio recordings of modern style music, there is little research that deals with automatic extraction of drum loops in polyphonic music audio. This paper presents an approach for drum loop pattern extraction, based on a technique of fusing the meter estimation information and onset clustering information. The extracted drum loop patterns are formed based on the temporal regularity (accent on meter) and drum types (bass and snare) of the beats. Despite not achieving very high-precision drum loop transcription, the extracted drum loop pattern characterizes the salient rhythm structure of the music, and can be very useful for categorization or similarity matching of music. Experimental results have shown the effectiveness of the proposed drum loop pattern extraction method.