Generator of Javanese Gamelan Musical Note Using 1D CNN Multi-Label Classification

A song in Javanese gamelan has various patterns in playing according to the event it accompanies; variations of this pattern are usually found on the accompaniment instrument, not on the song's main melody. Nevertheless, in Javanese gamelan, the note of the written melody is only the main melody, so new musicians need complete note for all instruments; one of them is bonang barung. Therefore, this research produces an automatic music generator for bonang barung instrument as a musical accompaniment from the multiple input balungan note, rhythm, and song structure. This research is one of the topics in Music Generation, where the proposed method used Multi-label Classification using 1D CNN. Meanwhile, compared with gamelan experts, the results of this automatic generator from bonang barung note have almost the same results because the accuracy value of the proposed method 1D CNN is 87.62%.