Pengenalan Nada Piano dengan Algoritma Short Time Fourier Transform (STFT)

Abstract

In the field of music, sheet music notation represents the graphical representation of the melody or harmony of a song. However, manually transcribing complex piano music can be challenging. In this research, we propose the application of Short-Time Fourier Transform (STFT) as a method for piano note recognition. STFT, a spectral analysis technique, is useful for analyzing frequency changes in time-varying signals such as music signals. The literature review reveals successful implementations of STFT in chord recognition and gamelan notation detection, with accuracies ranging from 60% to 90%. The research methodology includes a literature review, data collection of piano audio samples, feature extraction using Fast Fourier Transform (FFT), and system design involving preprocessing, segmenting the signal, feature extraction using STFT, signal processing using filters or thresholding, and mapping frequencies to piano notes. This research aims to provide an effective method for piano note recognition using STFT, contributing to automated music transcription and facilitating the learning and playing of piano music.

Keyword: Sheet music notation, Short-Time Fourier Transform (STFT), Piano note recognition, Fast Fourier Transform (FFT), Automated music transcription