Learning to play a gamelan instrument would be easier when there’s a musical notation guide. The process of converting a musical signal into a notation guide is called transcription. In this paper, we would like to transcript the gamelan music especially the Demung instrument using the Deep Learning method. Each Demung’s note from 6-low until 1-high would be converted to the time-frequency domain using STFT (Short-Time Fourier Transform). Then, those data will be treated as an input for the multilayers perceptron. The training method is a single label of each notation. The output returned by the model is a music roll transcription.