Recurrent Neural Networks for Drum Transcription

Music transcription is a core task in the field of music information retrieval. Transcribing the drum tracks of mu- sic pieces is a well-defined sub-task. The symbolic repre- sentation of a drum track contains much useful information about the piece, like meter, tempo, as well as various style and genre cues. This work introduces a novel approach for drum transcription using recurrent neural networks. We claim that recurrent neural networks can be trained to iden- tify the onsets of percussive instruments based on general properties of their sound. Different architectures of recur- rent neural networks are compared and evaluated using a well-known dataset. The outcomes are compared to results of a state-of-the-art approach on the same dataset. Further- more, the ability of the networks to generalize is demon- strated using a second, independent dataset. The exper- iments yield promising results: while F-measures higher than state-of-the-art results are achieved, the networks are capable of generalizing reasonably well.