KONVERSI NADA NADA AKUSTIK MENJADI CHORD MENGGUNAKAN PITCH CLASS PROFILE DAN NEURAL NETWORK BACKPROPAGATION

摘要

Chord recognition is a transcription of sound into chord, where audio input files will be classified according to different of provision level. Chord is several tones that sounded by musical instruments so that produce harmonic sounds. Chord which will be recognized in this application are major chord and minor chord. The process starts from wave sampling process. The next step is to divide the wave data into the frames (frame blocking). From frames are then transformed by Fast Fourier Transform (FFT). The process continued with the peak detection to retrieve peak values of the FFT. Frequency values that have peak amplitude are classified according to tune frequency use Pitch Class Profile (PCP). Results of PCP were used as input data on neural network backpropagation algorithm for training process. Keywords : Chord, Fast Fourier Transform, Pitch Class Profile, Backpropagation